Bibliography

Artworks and Exhibitions

He looks up.

 

Daniel Schraik is always looking up at the forest canopy. He observes how much light is falling onto the forest floor, and how much the trees are swaying in the wind. He needs the forest to be at an absolute standstill to get a sharp image. He is doing a ground truth measurement. 

 

The machine he is using to capture the forest does not have the fast shutter of a camera. This may not be a fair comparison, of course; the two apparatuses operate on entirely different mechanisms. 

 

Daniel stands next to a terrestrial laser scanner. Terrestrial laser scanning is a remote sensing measuring device that creates dense point clouds of the observed object. It is the state-of-the-art standard for vegetation monitoring. The machine emits a beam of laser towards the target, which then detects the reflected beam. Measuring the time needed for the laser beam to travel, the scanner computes the distance between the target and the object, thus creating an XYZ coordinate of one point of the scene. The machine repeats the process until it finishes scanning the whole scene. The end result is an assemblage of points, a point cloud, and one that represents 3D space. 

 

Each scan takes a couple of minutes to complete, so it is crucial that the forest is at a standstill; otherwise, the point cloud would be recording the movement rather than the space. The slightest movement creates a ‘smudge’ in the image: noise in the data.

 

I look straight ahead, with my camera pointing at him. I look at him, unaware of any breeze but definitely cognizant of the swarms of mosquitoes; it is a humid summer night. I work on a different timescale than Daniel does. I set my camera shutter to 1/60 second for this kind of wide portrait, sometimes even slower — to let as much light into the camera as possible in this dim environment. I know that if Daniel stands relatively still, the shutter speed should be enough to freeze him in the frame. I click while he scans.

 

Photographers, of course, care about image sharpness and details, and do everything to avoid blur. To a photographic process, blur is a source of noise. Ansel Adams, a household name in photography and a member of Group F64, devised the famous zone system to maximize tonal details (Adams 1948: 18). The zone system codifies the tonal range of a scene into eleven zones, from the darkest black to the brightest white. It provides a programmatic framework for photographers to consider the optimal exposure for a given scene, so that a photograph retains the maximum amount of detail across an extensive tonal range. The visual effect is achieved using a small aperture, allowing the resulting image to contain as much detail as possible. As a result of these methods, Group F64 became known for its immaculately crisp images of natural landscape1.

 

Daniel waits for the wind to stop; I adjust my aperture.

 

Artists are celebrated, it’s because of a particularly innovative ways of creating images, of depicting the human condition. Their works are hung on museum walls for viewers to gaze upon, to ponder, to stand in front of pensively. Leonardo DaVinci, Pablo Picasso, Mark Rothko, Agnes Martin, Susan Meiselas. But artists do not have the monopoly on image-making; they never have. Scientists have been making technical images alongside artists for centuries. Their images have equal cultural significance, and sometimes more. Scientific botanical illustrations may not follow John Ruskin's artistic vision of rendering nature scenery, but they do depict accurate plant anatomy in vibrant colours and vivid details. Their nature images are also less widely acknowledged than those of some well-known artists — Anna Atkins’ botanical cyanotypes, for instance, or Karl Blossfeldt’s macrophotography images of plants. Yet, in fact, the categorical divide between artistic and scientific image-making may be as purely ideological as it is unnecessary. Taking a small step outside these artificial dichotomies, one can see that Ansel Adam’s zone system was arguably as much algorithmic as it was aesthetic. The invention of the zone system was underpinned by several factors: Adams’ attention to craftsmanship, his pursuit of beauty, and his environmentalist philosophy. But it was also motivated simply by his desire to ‘[secure] the appropriate information on the negative’ (Adams 1948: iv). In this sense, the zone system has never been just a creative tool by which artists could produce engrossing landscape photographs. It has always also been an algorithmic scheme for preserving the maximum amount of information of any real-life scenery converted into placement on flat, photosensitive surfaces. The pursuit of informational accuracy and visual aesthetics have never been mutually exclusive.

 

Much advancement in image-making is motivated by scientific curiosity. From telescopes to microscopes, terrestrial laser scanners to the Event Horizon Telescope, scientists have made images of the infinitely big to the infinitesimally small, arresting nature, whether visible or invisible, near or far in front of our eyes.  Writing about Harold Edgerton's high-speed photographs of water splashes and bullets frozen in mid-air, James R. Killian expressed his amazement at photography's ability to make the invisible visible. (Edgerton and Killian 1954) Edgerton is not the first person to freeze a water splash. Working at the dawn of modern photography, British physicist Arthur Worthington attempted to freeze a water splash with his own eyes (Worthington 1895). He devised a spark-producing apparatus that flashes at the moment the water drop hits the surface. In a dim room, he would observe the water splash in pulses of light and record the shape of the splash that was imprinted onto his eye at the split second. He abandoned his drawing practice after discovering the modern camera. Upon looking at the photographs, he realized how wrong his drawings were. The perfect symmetrical form of water splashes he thought he saw was indeed irregular. In his confession, he praised the "objective view" of the camera.


Since its inception, writers have given different accounts to explain the objectivity of the camera. William Fox Talbot wrote about automatic drawing (Talbot 1844). John Szarkowski remarks that a photograph has a unique physical connection to reality, as it records the imprint of light patterns cast by the objects in front of the lens (Szarkowski 2012) . Others emphasize the unrivalled detail of a photo print. Many early photography accounts in the nineteenth and twentieth century emphasize its automatism and proximity to nature, often brushing over human intervention in the photographic process. Talbot was famously known for his botanical 'sun drawings'. However, at the same period, as he was developing the calotype, he also discovered the "delight" of artistically arranged pictures. Talbot experimented with staging photographs with some servants posing as fruit sellers in his estate in Lacock Abbey (Talbot & Jones 1845)2. The French artist Hippolyte Bayard made a staged self-portrait in which he posed as a drowned man in 1840 – one of the many early examples of creative uses of photography.


With my camera pointing at Daniel, I asked him to stop what he is doing and hold his pose for a while. I did not stage the photograph, but the photograph is not objective either. I frame, I compose, I intervene, but my intervention does not make what is on the photograph a complete lie. Image is always a partial truth and a partial illusion. Different paradigms have their own way of validating the truth within the image while normalizing its artifice. Philosopher Louis Althusser remarked that while images constitute an illusion, "they make allusion to reality and they need only be 'interpreted' to discover the reality of the world behind their imaginary representation of that world." (Althusser 1999: 320) The realism in an image is supported by an ideology. As Althusser may phrase it, images are the material manifestation of the ideological apparatus, representing the imaginary relationship between individuals and their real conditions of existence.

On this particular field trip to the Hyytiälä field station, I took several landscape photographs. Landscapes are pictures that depict natural scenery. Early photography theorists considered photographs to be an anomaly in the pictorial tradition because they do not merely capture Nature, as in drawing; rather, as Daguerre claimed, photographs "give her the power to reproduce herself" (Batchen 1999:66). The realism of a landscape is sustained by the ideology that the photograph replicates nature with minimal human intervention — that a piece of landscape photography is nature itself. 

 

Reconfiguring this imaginary relationship is central to the arguments of writers like Geoffrey Batchen and Rosalind Krauss, who have instead argued that landscape is mediated nature rather than pure natural view (Batchen 1999, Krauss 1982). Batchen, for instance, has noted that concurrent to the invention of photography, landscape was considered under the aesthetic regime of the picturesque. The picturesque put forth rules about framing and composition "expressive of that peculiar kind of beauty which is agreeable in a picture", forming a technical language and conceptual framework by which nature is considered to be translated onto an image (Batchen 1999: 72). 

 

From early photography to today, landscape — even with its aim centering on the capturing of nature — is neither objective nor natural. Rather it embodies the nature-culture, ideal-real paradox that underpins visual representation. According to Batchen, the paradox is apparent as early as the nineteenth century; it can be found in Talbot’s writings, among others. Talbot wrote that photography is both a mode of drawing and a system of representation in which no drawing takes place. His inability to pin down a singular definition hinted at the multifaceted nature of photography as something simultaneously both natural and cultural.

 

Daniel and I are in nature together, but we are creating different landscapes at the same time. Holding a camera, I, perhaps subconsciously, direct my eyes in the manner of the picturesque, concentrating on composing with the zone system. Daniel, operating a laser scanner, is concerned with gathering accurate data for the development of a model.

 

Objectivity assumes a view from nowhere, or a neutral apparatus from which the world can be observed. This imaginary relationship between the world and the individual is an ontology — one that suggests that the observing subject is independent of the observed, and that the acts of observation and use of the apparatus through which one observes do not affect reality. Images are the material existence that sustains this ideology of objectivity. People have been attempting the invention of such machines that produce ‘objective’ images for centuries. Worthington and others in the history of photography praised the camera for its objectivity. But in truth, machines are not objective. Machines measure. They are simply manifestations of human intention. Human intention is what decides their programming and their uses. 

 

The portrait I am making of Daniel is not nature itself. It will always contain my intention to communicate my experience. It is not a sign, but a signpost pointing towards the world. Photographers such as Worthington believed the camera was objective — that its images allowed for a truer depiction of reality, one that could counteract the flaws of the human eye. But technical vision does not guarantee better vision. It seems so only when people abiding by a representational system accept its artifices; the adjusting of exposure is standard, the adding of colour filters is acceptable.

 

At the same time, other inventions are considered to go too far, to interfere with realism. Digital cloning is one, for instance — considered a violation of the realism of a photograph. These rules are historical and contingent, constantly changing according to current trends and technological developments. Photography scholar John Tagg remarked that "realism is a social practice of representation, an overall form of discursive production, a normality which allows a strictly delimited range of variations" (Tagg 1999: 271). To continue his line of thought under this specific context, realism is the process in which what is really culture is imagined to become nature. It is a discursive process in which a cultural artefact becomes conflated with nature itself, an imaginary relationship naturalized and institutionalized. Nineteenth-century realist paintings, the Tank-Man photograph, the digital image of the M87 black hole created using data captured by the Event Horizon Telescope — are all ‘real’. Still, their realisms are sustained within different epistemological paradigms. An image is a symptom of an ideological apparatus that configures ‘the imaginary’ and ‘reality’. A photograph is one manifestation; a point cloud is another. As Daniel and I engage in an interdisciplinary dialogue, we look back into the parallel apparatuses we operate, and thinking through each of our entanglements between image and reality. 

 

*******

 

On this summer night, Daniel and I are standing on a forest plot at Hyytiälä field station in the middle of Finland. We are both observing from the ground level, but for different reasons. Daniel belongs to a team of remote sensing researchers led by associate professor Miina Rautiainen. The official aim of their research project, ‘From needles to landscapes: a novel approach to scaling forest spectra’, is to develop a better model for the interpretation of multi-sensor satellite images of forests (Hovi et al 2019). Remote sensing is the technique of near real-time environmental monitoring using open access data without on-site observation (Schraik 2022: 15). Remote sensing frequently refers to planetary observation. For instance, the Opportunity Rover, equipped with a panoramic camera capable of photographing a scene with thirteen colour filters, was designed as a remote sensing endeavour to explore the extra-terrestrial landscape of Mars. On Earth, orbiting satellites collect data on environmental elements from surface temperature to vegetation coverage through multispectral sensors. The network of satellites forms a planetary sensorial system monitoring Earth in real-time. Each satellite imagery is telepresence: relaying the top-down view once imagined as a supernatural power, now becomes accessible to the human eye.

 

Photography is also a form of telepresence, but one mostly confined to the visible light spectrum. Media theorist Marshall McLuhan described media as an extension of the human senses (McLuhan 1964). In that sense, a photograph is a tool able to transport perspective from a single position, be it geographical, socioeconomic or political, to another. Scientific photography, space photography, remote sensing all do this uniquely.

 

The Sentinel satellites, for instance, record images of multiple spectrums — from high-resolution optical photographs to radar and infrared imaging. The satellite’s gaze dissects the Earth into multiple electromagnetic spectra, each with its own visibilities and invisibilities. Optical light is natural to the human eye but is easily obstructed by the Earth’s atmosphere; Infrared light is sensitive to photosynthetic surfaces but has poor spatial resolution. Researchers switch between different data layers, triangulating useful information from the complementary spectrums. 

 

I like to think of remote sensing as vision escaping the human body. The sensorial organ expands to a distributed global network, and perception escapes the visual cortex to be incorporated into algorithms. The information contained in an image is no longer confined to the humanly perceptible, but now, to the machine-readable (Bratton 2015, Hansen 2015).

 

The portrait I have taken of Daniel, I have scanned at 3200ppi in 16 bits RGB, occupying 360.8 MB of memory. The image on the negative is 68mm by 56mm. The digital scan is a matrix of 8540 by 7040 pixels. Each pixel contains information that expresses one of 65,536 possible tonal variations on an RGB scale. The colour depth gives an illusion of smooth tones even though the image comprises discrete squares of colours. 

 

The satellite image taken by Sentinel 2 using a multispectral instrument on 7 May 2021 is 665.96 MB. The image, covering an area of 100 square kilometres, contains the reflectance map of multiple spectral bands. Each map records the intensity of a certain wave spectrum as reflected back from Earth’s surface to the satellite. Some spectrums are invisible to the human eye. Each image contains deep layers of information. The information of a map can be translated to a value in one of each RGB channel, which, when combined, form a colour image. This type of image is also called a false colour image because the optical layer is an interpretation of the data rather than the physical spectrum registered by the sensors. The red in a false colour image does not have a direct physical relationship with the world as we experience the colour red; here it signifies the reflectance of an invisible wave. 

 

In these practices, a camera is a measuring apparatus. A photograph depicts the geometric measurement of visible light in a given direction for each pixel (Yiu and Schraik 2022). From the scientist’s point of view, the informational capability of a photograph eclipses its aesthetic quality. This view is further exemplified by the digitization of the medium and the advent of novel computational techniques. Artist/theorists such as Harun Farocki also investigate the varied social roles of images, as in his well-known trilogy Eye/Machine that examined the military-industrial complex (Farocki 2003). Farocki’s notion of ‘operational image’ succinctly points out the new functionalities of photography beyond mere representation, drawing attention to fields in which images are embedded into a computer vision system and used as interfaces to simulate combat and target moving objects. Farocki’s works are frequently referenced in a new wave of exploratory thinking that shifts the focus from the visual to the unvisual, work that scrutinizes the computational dimension (Ingrid and Hoelzl 2015, Beller 2016); and also, its underlying networked infrastructure (Sluis and Rubinstein 2008, MacKenzie and Munster 2019).

 

These practices stand in stark contrast to traditional, narrow, optical, or parochial views of photography. (Rubinstein 2020: 5). Contrary to contemplation of the individual aesthetic piece, they address images from a systemic perspective. They theorize the image not as the representation of reality, but as ‘a button with a picture’ (Bratton 2015: 224). Those who work in such areas are deliberately attempting to rid themselves of the ocular-centric, the perspectival, and the representational baggage. Such work constitutes what is often known as post-representational photography theory (Rubinstein 2018: 8-18).

 

Daniel describes a pixel as a coordinate on an image, from which he is able to manipulate and extract information. Thus a pixel serves as a repository for data. Data, while it encodes information, also stands in for our absence, and thus becomes a proxy of the world. Satellite images, for instance, are signs of the presence of data and the absence of the human witness. Most satellite images depict wavelengths invisible to humans, collected by non-human agents. executing pre-written programmes. How can one justify the sporadic absence of humans in the process of information acquisition? 

 

This is a question that is equally relevant to photography, even as there the absence of the human may be less obvious given that photographers maintain more direct contact with the camera. At the same time, my finger on the shutter does not give me more control over the internal automatic camera mechanism. The camera remains a black box. Articulating a new paradigm in photography, Daniel Rubinstein has recently written that "[w]hether a camera or computer, a black box is a device with an input and an output. If you feed data into a black box, it will be output as information" (Rubinstein 2020: 4). No matter how thoroughly a photograph imitates the human visual experience of the world, it does not disguise the fact that the photograph is a stand-in for something the viewer did not experience physically. Photography is, thus, remote sensing in the broadest sense. Both mediums require an imaginary relationship between the image and reality to justify the human absence. In both cases, human presence can be either emphasized or minimized in the creation of respective realisms.

On a summer night, I took a photograph of Daniel conducting a laser scanning measurement on the forest floor. I was with him; I witnessed his fieldwork in person. The mosquito bites on my legs were my proof. The pixels on the digital scan of his portrait suggest the same evidence of the same event. The difference is that the photograph lasts longer; the bites fade away. The photograph also exists outside my body, becoming a proxy of my existence. 

 

There are many reasons people tend to believe in photography: the sharpness, the details, the physical imprint of lights, the similarities to human visual experience, the trust in the photographer, the reputation of the newspaper in which the photo is published. However, the advent of digital photography in the 1990s has in fact added another layer of artifice to photographic realism. Art historian Geoffrey Batchen has rightfully pointed out that in the digital image, the referent in a visual work is no longer the physical world but "differential circuits and abstracted data banks of information" (Batchen 2002: 140). Photography becomes programmable, networked, and non-human (Beller 2016, Rubinstein and Sluis 2008, Zylinska 2017). Each pixel contains not physical light as in earlier stages of photography, but encoded data. In this sense, a digital photograph is really a false colour image optimized for the human eye. It does not share a direct causal relationship with the human perception of the physical world. 

 

Despite its artifice being frequently pointed and from multiple directions, photography nonetheless remains a dominant social medium of truth and evidence that something has occurred in reality. In a speech he gave at the Sorbonne on the occasion of celebrating the centenary of photography on 7 January 1939, French poet Paul Valery reminded of pre-photographic time when facts were established simply by sufficient numbers of people testifying as to their observations. Within the century, photography eroded these channels of knowledge, with one snapshot proof enough to destroy the testimony of some hundred people who had seen an event first-hand. (Trachtenberg 1980: 196). 

 

Photography's authority in relation to ‘truth’ had been foreshadowed precisely one hundred years earlier, in a different speech, given by a different Frenchman theorizing on the image. In 1839, Louis Daguerre presented his invention at the Academy of Science in Paris. He asserted the uniqueness of the daguerreotype, an early prototype of modern photography, in its ability to empower nature to reproduce itself (ibid: 66). Daguerre’s characterized for the audience the photographic process as automatic physical and chemical imprint capable of producing stunning details, without any idea of drawing and unalterable by light. This explanation maintained its primacy through evidentiary power for decades to come. Through this same period, the ‘truth claim’ of photography has persisted throughout its history, often challenged but never overturned. As photography moved from the physical-chemically based at its birth to its data-based form in the digital era, the truth claim has shifted in reasoning, but not in assertion. One thing that has remained the same is that as reality has been ever increasingly mediated, the medium exerts power over the naked eye as to what ‘really’ happened. In a perpetually-mediated world, the photographic image is taken as ground truth.

 

Valery raised a good point about the nature of truth, reality, and evidence in his 1930s speech. In the most original sense, ‘ground truth’ means reality as experienced first-hand rather than by report. By definition, ground truth is embodied and non-transferable. But as photography has become irreversibly interwoven into military surveillance, scientific exploration, and the judiciary system as the proxy of truth, photographs have become ground truth. Such photographs, intended to communicate reality, are referred to as documentaries. When Susan Meiselas photographed a man throwing a Molotov cocktail at the Nicaraguan National Guard fortress in 1979, she was mediating that image through her first-hand experience of the Nicaraguan Revolution as a photojournalist (Meiselas 1979). The iconic photograph, known as the Molotov Man, could have been staged — photography has always been susceptible to manipulation since the medium's inception even before the digital era. But underpinned by the photo-documentary tradition, a suspension of disbelief occurs. The viewer ignores the possible artifice and begins to believe in the photograph's realism; they see it and so it must be true. To take a photograph as ground truth is to accept the imaginary relationship between the image and reality.

 

The term ground truth did not come up in my own mind when I took the photograph of Daniel; rather, he mentioned it while explaining the history of his discipline to me. In remote sensing, ground truth refers to data collected on-site, which is then used to calibrate, to build models, to decipher information from images, or to spot a tree from a pixel. Yet even such data has its limitations. Despite state-of-the-art image sensors, infrared and near-infrared imaging, due to the nature of waveband, is limited by spatial resolution. Longer wavelengths, which are useful in detecting foliage, tend to have lower energy. This means that a pixel needs to cover a larger area than the optical wavelengths it represents, if it is to receive enough energy for a satellite sensor to register any useful activity. For certain wavebands, a pixel may need to cover dozens of square meters of land. 

 

On a human scale, a pixel is much smaller than the object that it represents. A common modern digital display has over 200 pixels packed in every inch; each side of the pixel is smaller than one-tenth of a millimetre. On a planetary scale, spatial resolution breaks illusion. Objects are hidden underneath pixels. To ‘see’ a tree through a satellite image is analogous to counting skin cells from a portrait. At resolution limits, seeing as a way of knowing breaks down. And the interpretation of data becomes that much more crucial. 

 

Daniel collects ground truth data by documenting the spatial topology of the forest using terrestrial laser scanning. Comparing ground truth data with satellite data, his team establishes their mathematical correlation – a model that deciphers the forest landscape underneath the pixel. When the scanning finished, Daniel shows me the resulting image on the scanner's viewer. The collection of red dots of different shades depicts trees and the undergrowth of the forest floor. I understand immediately that I am looking at a 3D representation of the forest. A point cloud depicts the forest as a constellation of countless discrete points, as if seeing through the eyes of an extraterrestrial species. For Daniel, the point cloud is a raw material for building a better interpretation model. I am looking at an image; he is building a model. I see it as the end; he sees it as a means to an end. Both point toward the reality we share, oscillating on three realisms – experience as ground truth, images as ground truth, data as ground truth.

 

Daniel's point cloud, the laser scanner that produces it, the field trip we took to capture it and the research funding that supports it belongs to a social practice of representation I refer to as data realism. Different from photographic representation, however, this form of realism posits that the relationships between data formulated through mathematical operations represent our lived reality. Data realism is sustained by computational models. The common aphorism goes, 'all models are wrong, but some are useful.' A useful model has predictive power over reality. Models improve with more data and rigorous measurements. The point cloud of the forest is only one piece of the puzzle for the model called the reflectance model. By measuring the spatial topology of the forest, Daniel aims to formulate its mathematical correlation with the satellite's multispectral data. His colleagues meticulously measure the spectral properties of tree shoots under a goniometer to understand the way leaves reflect a certain band of electromagnetic waves.3 These measurements are recorded not only in photographs but also as point clouds, spectral graphs, and bitmap scans. And then the point cloud of the forest is only one piece of the puzzle for the model called the reflectance model. In fact, thousands of data points make up a reflectance model — creating a mathematical description of the interaction between sunlight and tree.4

In data realism, models become nature itself. Unable to contain the full complexity of the world, statistical models are prone to uncertainty. Uncertainty can be somewhat tamed with sufficient data and reiterative computation, such that models can create accurate estimation of reality: close enough to factual reality to be considered good enough to justify the human absence in the act of observation.

 

Data realism is further sustained by social networks and neural networks. Global network infrastructure facilitates the collection of data, while machine learning refines reiteration computation. Like polishing a mirror, remote sensing constantly tries to use data to improve models to better reflect reality. However, if Batchen was right about the visual no longer representing the real world but merely representing data, then that mirror no longer signifies an unmediated reality for which photography strove for centuries. Instead, it has become a metaphor for "a dynamic enfoldment of opposites, a movement that incorporates without synthesizing the conceptual poles nature-culture, real-ideal, general-particular, science-art, object-subject, reflection-expression" (Batchen 1999: 81).

Daniel and I are looking in opposite directions. I look forward, with my optical lens. I look through the viewfinder at him. I press the shutter, already having a good idea as to how the photograph will turn out. Daniel looks backwards, with his model. He began with the data from a satellite image. Using a reflectance model and ground truth data, he works backwards to decipher the original scene, that creates the image we are both able to take in different forms. 

 

I embalm our presence; he resurrects a tree.

All images besides the author’s own artwork are derived from two sources. Some are archival images from the Finnish Forest Museum acquired by the author. Other images are part of the dataset generously shared by a remote sensing research group based at Aalto University (‘From needles to landscapes: a novel approach to scaling forest spectra’). This data is licensed according to the open science principle and under the Creative Commons license CC-BY 4.0. 

Adams, Ansel, The Negative: Exposure and Development. Ansel Adams Basic Photography Series/Book 2 (Boston: New York Graphic Society, 1948)

Althusser, Louis, ‘Ideology and Ideological State Apparatuses’, in Visual Culture: The Reader, ed. by Jessica Evans and Stuart Hall (Thousand Oaks: SAGE, 1999)

Batchen, Geoffrey, Burning with Desire: The Conception of Photography, (Cambridge: MIT Press, 1999)

Batchen, Geoffrey, Each Wild Idea: Writing, Photography, History, (Cambridge: MIT Press, 2002)

Beller, Jonathan, ‘The Programmable Image of Capital: M-I-C-I′-M′ and the World Computer’, in Postmodern Culture, 26.2 (2016), https://doi.org/10.1353/pmc.2016.0005

Bratton, Benjamin H., The Stack: On Software and Sovereignty, Software Studies (Cambridge: MIT Press, 2015)

Edgerton, Harold Eugene, and James Rhyne Killian. Flash!: Seeing the Unseen by Ultra High-Speed Photography. (Michigan: C.T. Branford, 1954) 

Hansen, Mark B. N., Feed-Forward: On the Future of Twenty-First-Century Media (Chicago: University of Chicago Press, 2015) 

Hoelzl, Ingrid, and Rémi Marie, Softimage: Towards a New Theory of the Digital Image (Bristol: Intellect, 2015)

Hovi, Aarne, Eva Lindberg, Mait Lang, Tauri Arumäe, Jussi Peuhkurinen, Sanna Sirparanta, et al., ‘Seasonal Dynamics of Albedo across European Boreal Forests: Analysis of MODIS Albedo and Structural Metrics from Airborne LiDAR’, in Remote Sensing of Environment, 224 (2019), 365–81 https://doi.org/10.1016/j.rse.2019.02.001

Krauss, Rosalind, ‘Photography’s Discursive Spaces: Landscape/View’, in Art Journal, 42.4 (1982), 311–19 https://doi.org/10.1080/00043249.1982.10792816

MacKenzie, Adrian, and Anna Munster. ‘Platform Seeing: Image Ensembles and their Invisualities’, in Theory, Culture & Society, 36.5 (Sept. 2019)

McLuhan, Marshall, Understanding Media: The Extensions of Man (Cambridge: MIT Press, 1994)

Rubinstein, Daniel, and Katrina Sluis, ‘A Life More Photographic: Mapping the Networked Image’, in Photographies, 1.1 (2008), 9–28 https://doi.org/10.1080/17540760701785842

Rubinstein, Daniel, ‘Post-representational Photography, or the Grin of Schrödinger’s Cat’ in Photography Reframed: New Visions in Contemporary Photographic Culture ed. by Benedict Burbridge and Annebella Pollen (London New York: I.B. Tauris, 2018), 8-18

Schraik, Daniel, Clumping in Forest Radiation Regime Models (doctoral thesis, Aalto University, 2022)

Sontag, Susan, On Photography (New York: Rosetta Books, 2005)

Szarkowski, John, The Photographer's Eye (New York: The Museum of Modern Art, 2012)

Tagg, John, ‘Evidence, truth and order: a means of surveillance’ in Visual Culture: The Reader, ed.by Jessica Evans and Stuart Hall (London ; Thousand Oaks: SAGE Publications in association with the Open University, 1999), 244-273

Talbot, William Henry Fox, The Pencil of Nature (Longmans: London, 1844), http://www.gutenberg.org/files/33447/33447-h/33447-h.html

Valéry, Paul, ‘The Centenary of Photography’, in Classic Essays on Photography, ed. by Alan Trachtenberg (New Haven: Leete’s Island, 1980)

Worthington, A. M., Splash of a Drop (London: Society for Promoting Christian Knowledge, 1895)

Zylinska, Joanna, Nonhuman Photography (Cambridge: MIT Press, 2017)

Farocki, Harun. Eye/Machine Trilogy. Video, 2003

Yiu, Sheung and Daniel Schraik, ‘How To See Something Where There Is Nothing (With a Remote Sensing Researcher)’ Exhibition, Titanik, Turku, Finland, 2022

Talbot, William Henry Fox and Calvert Richard Jones, The Fruit Sellers. Photograph, 1845

Meiselas, Susan, Molotov Man. Chromogenic print, 1979, Princeton University Art Museum

Sheung Yiu

Spotting A Tree From A Pixel (With Remote Sensing Researchers)

Ground Truth [n.]


  1. In everyday life, the reality of a situation as experienced first-hand rather than by report.

  2. In remote sensing, information collected on location. Information that then allows image data to be related to real features and materials on the ground.

  3. In statistics and machine learning, ground truth means checking the results of machine learning for accuracy against human-drawn results.

  4. In remote sensing, actual data as ascertainable through direct observation rather than through inference.

5.57006979 10.73803997 -1.08080995 254 247 255 0.311589
5.59301996 10.78246975 -1.08130002 254 247 255 0.291569
5.63402987 10.86157036 -1.08130002 254 247 255 0.290349
5.67406988 10.93920994 -1.07349002 254 247 255 0.342611
5.69408989 10.97681046 -1.07349002 254 247 255 0.312810
5.74047995 11.06665039 -1.07446003 254 247 255 0.295720
5.75414991 11.09302044 -1.07299995 254 247 255 0.310369
5.77319002 11.12963963 -1.07251000 254 247 255 0.335775
5.78539991 11.15258980 -1.07104003 254 247 255 0.358480
5.79174995 11.16528034 -1.06812000 254 247 255 0.425147
5.90502977 11.38354015 -1.07251000 254 247 255 0.282048
5.92260981 11.41724014 -1.07202005 254 247 255 0.330144
5.93335009 11.43871975 -1.07007003 254 247 255 0.352865
6.14037991 11.83813000 -1.08178997 254 247 255 0.329900
6.15160990 11.85962009 -1.07983005 254 247 255 0.374594
6.16137981 11.87817001 -1.07738996 254 247 255 0.372396
6.17309999 11.90112019 -1.07494998 254 247 255 0.329168
6.29662991 12.13842964 -1.07934999 254 247 255 0.321355
6.44652987 12.42846966 -1.08715999 255 247 255 0.299138
6.48853016 12.50854015 -1.09009004 255 247 255 0.304021
6.51245022 12.55492973 -1.08959997 255 247 255 0.296696
6.51489019 12.56029987 -1.07202005 255 247 255 0.111345
6.62817001 12.77758980 -1.08617997 255 247 255 0.227832
6.67603016 12.86987019 -1.08959997 255 247 255 0.205127
6.72533989 12.96508980 -1.09301996 255 247 255 0.178012
6.76635981 13.04419041 -1.09496999 255 247 255 0.160189
6.76635981 13.04419041 -1.09057999 255 247 255 0.141390
6.81420994 13.13696003 -1.09351003 255 247 255 0.139193
6.95190001 13.40209961 -1.11059999 255 247 255 0.142367
7.18529987 13.85181046 -1.12328994 255 248 255 0.164828
7.22437000 13.92749023 -1.12476003 255 248 255 0.254688
7.53979015 14.53540039 -1.12720001 255 248 255 0.174838
7.57934999 14.61205959 -1.12817001 255 248 255 0.203906
7.56127977 14.57738972 -1.12036002 255 248 255 0.188510
7.54760981 14.55053997 -1.11303997 255 248 255 0.204395
7.55541992 14.56616020 -1.10913002 255 248 255 0.198276
7.59448004 14.64087009 -1.10961998 255 248 255 0.190951
7.61156988 14.67407036 -1.10718000 255 248 255 0.192416
7.62085009 14.69213963 -1.10327005 255 248 255 0.161410
7.60571003 14.66283989 -1.09546006 255 248 255 0.180941
7.70922995 14.86254978 -1.10521996 255 248 255 0.255177
7.76049995 14.96068954 -1.10718000 255 248 255 0.288640
7.82837009 15.09204006 -1.11157000 255 248 255 0.211963
7.88891983 15.20874023 -1.11499000 255 248 255 0.165560
7.92456007 15.27758980 -1.11450005 255 248 255 0.131853
7.31518984 14.10278034 -1.02367997 255 248 255 0.091081
7.99389982 15.41137981 -1.11353004 255 248 255 0.188266
8.03003025 15.48071003 -1.11303997 255 248 255 0.136995
8.07056046 15.55834961 -1.11303997 255 248 255 0.134066
8.16870022 15.74829006 -1.12134004 255 248 255 0.151400
8.28345013 15.96947956 -1.13110006 255 248 255 0.265431
8.32104015 16.04125977 -1.13062000 255 248 255 0.288396
8.35424995 16.10570908 -1.12915003 255 248 255 0.285466
8.37182999 16.13989067 -1.12621999 255 248 255 0.311589
8.38939953 16.17407036 -1.12279999 255 248 255 0.288640
8.44896984 16.28833008 -1.12476003 255 248 255 0.234424
8.45629978 16.30249023 -1.12036002 255 248 255 0.386801
8.48803997 16.36352921 -1.11840999 255 248 255 0.305241
8.55688000 16.49633980 -1.12182999 255 248 255 0.131853
8.78442001 16.93530083 -1.14575005 255 248 255 0.181674
8.81470013 16.99341011 -1.14380002 255 248 255 0.220508
8.84204006 17.04663086 -1.14136004 255 248 255 0.230274
8.93920994 17.23315048 -1.14136004 255 248 255 0.307683
8.94896984 17.25268936 -1.13646996 255 248 255 0.326970
8.95435047 17.26293945 -1.13110006 255 248 255 0.264698
8.98021984 17.31225967 -1.12817001 255 248 255 0.268376
9.00757027 17.36548042 -1.12573004 255 248 255 0.254688
9.05688000 17.46068954 -1.11938000 255 248 255 0.247364
9.06665039 17.47924995 -1.11450005 255 248 255 0.210498
9.07544041 17.49633980 -1.10961998 255 248 255 0.180453
9.27807999 17.88695908 -1.12768996 255 248 255 0.146029
9.38500977 18.09350967 -1.13452005 255 248 255 0.337728
9.40845013 18.13891983 -1.13062000 255 248 255 0.322576
9.46850967 18.25366020 -1.13159001 255 248 255 0.328923
9.52807999 18.36938095 -1.13207996 255 248 255 0.297185
9.53588963 18.38451958 -1.12671006 255 248 255 0.229053
9.57007027 18.44945908 -1.12426996 255 248 255 0.182895
9.68579006 18.67309952 -1.13110006 255 248 255 0.224414
9.74682999 18.79125977 -1.13159001 255 248 255 0.263233
9.95238972 19.18676949 -1.11450005 255 248 255 0.263233
10.10132027 19.47339058 -1.09692001 255 248 255 0.254444
10.14867973 19.56469917 -1.09496999 255 248 255 0.275944
10.20386028 19.67163086 -1.09398997 255 248 255 0.281804
10.26537991 19.79027939 -1.09398997 255 248 255 0.305241
10.33129978 19.91772079 -1.09398997 255 248 255 0.313054
10.37182999 19.99486923 -1.09106004 255 248 255 0.307683
10.41528034 20.07885933 -1.08861995 255 248 255 0.283513
10.50463963 20.25171089 -1.09057999 255 248 255 0.246387
10.55883980 20.35668945 -1.08911002 255 248 255 0.252735
10.62084961 20.47583008 -1.08813000 255 248 255 0.239063
10.68091011 20.59155083 -1.08715999 255 248 255 0.240772
10.72192001 20.67016983 -1.08374000 255 248 255 0.225635
10.77954006 20.78149033 -1.08228004 255 247 255 0.218799
10.83129978 20.88109970 -1.08032000 255 247 255 0.238331
10.84887981 20.91527939 -1.07446003 255 247 255 0.201709
10.89185047 20.99877930 -1.07152998 255 247 255 0.203662
11.14478016 21.48608017 -1.04321003 255 247 255 0.183627
11.19946003 21.59106064 -1.03296006 255 247 255 0.098650
11.67650986 22.51099014 -1.06128001 255 247 255 0.187289
11.79858017 22.74633980 -1.05639994 255 247 255 0.125017
11.95044041 23.03931046 -1.06176996 255 247 255 0.156283
11.99779987 23.13013077 -1.05736995 255 247 255 0.243946
12.03295994 23.19799995 -1.05249000 255 247 255 0.265431
12.05541992 23.24146080 -1.04663002 255 247 255 0.241993
12.13599014 23.39673042 -1.04516995 255 247 255 0.236866
12.18432999 23.48999023 -1.04077005 255 247 255 0.250050
12.23608017 23.59008980 -1.03687000 255 247 255 0.205615
12.60766983 24.30640030 -1.04271996 255 247 255 0.205615
12.63599014 24.36059952 -1.03638005 255 247 255 0.289372
12.65845013 24.40453911 -1.03003001 255 247 255 0.243702
12.67504978 24.43578911 -1.02222002 255 247 255 0.244434
13.02367973 25.10766983 -1.02417004 255 247 255 0.171664
13.03637981 25.13306046 -1.01586998 255 247 255 0.162631
13.14136028 25.33521080 -1.01538002 255 247 255 0.114519
13.36548042 25.76684952 -1.02319002 255 247 255 0.120134
13.50366020 26.03296089 -1.02514994 255 247 255 0.144564
13.55932999 26.14086914 -1.01978004 255 247 255 0.169223
13.60620022 26.23168945 -1.01440001 255 247 255 0.135286
13.64525986 26.30591011 -1.00757003 255 247 255 0.124285
13.86548042 26.73119926 -1.01488996 255 247 255 0.204395
13.91283989 26.82201958 -1.00854003 255 247 255 0.207080
13.97290039 26.93774033 -1.00365996 255 247 255 0.178012
14.09204006 27.16772079 -1.00220001 255 247 255 0.219043
14.13599014 27.25268936 -0.99585003 255 247 255 0.213184
14.18237019 27.34203911 -0.98949999 255 247 255 0.204639
14.52221966 27.99682999 -0.97387999 255 247 255 0.245655
14.53442001 28.02025986 -0.96460003 255 247 255 0.225391
14.53979015 28.03100967 -0.95482999 255 247 255 0.219043
14.54467964 28.04076958 -0.94555998 255 247 255 0.278874
14.54516983 28.04174995 -0.93579000 255 247 255 0.294011
14.54761028 28.04565048 -0.92602998 255 247 255 0.223438
14.54955959 28.04956055 -0.91626000 255 247 255 0.250050
14.55103016 28.05298042 -0.90649003 255 247 255 0.236133
14.55249023 28.05591011 -0.89624000 255 247 255 0.191440
14.64966011 28.24291992 -0.85277998 255 247 255 0.277409
14.66137981 28.26538086 -0.84302002 255 247 255 0.269841
14.66431046 28.27075005 -0.83374000 255 247 255 0.277409
14.66724014 28.27660942 -0.82397002 255 247 255 0.303777
14.66967964 28.28149033 -0.81371999 255 247 255 0.274968
14.67212009 28.28638077 -0.80396003 255 247 255 0.256153
14.67407036 28.28931046 -0.79418999 255 247 255 0.275944
14.67603016 28.29369926 -0.78394002 255 247 255 0.315984
14.67700005 28.29565048 -0.77465999 255 247 255 0.336019
14.67992973 28.30150986 -0.76440001 255 247 255 0.259083
20.85717964 40.21020889 -0.90161002 255 247 255 0.206104
20.91724014 40.32543945 -0.88989002 255 247 255 0.168246
21.06176949 40.60425186 -0.88208002 255 247 255 0.072526
21.35961914 41.17895889 -0.87963998 255 247 255 0.140169
21.89429092 42.20922852 -0.85718000 255 247 255 0.108659
22.01147079 42.43529892 -0.84692001 255 247 255 0.104753
20.34057999 39.21361923 -0.67211998 255 247 255 0.016846
20.33861923 39.21020889 -0.64428997 255 247 255 0.018067
20.33275986 39.19945908 -0.63012999 255 247 255 0.020020
20.33179092 39.19702148 -0.61646003 255 247 255 0.021241
20.33081055 39.19556046 -0.60277998 255 247 255 0.024414
20.32983017 39.19311905 -0.58862001 255 247 255 0.021241
20.33081055 39.19506836 -0.57494998 255 247 255 0.016846
20.32836914 39.19067001 -0.56079000 255 247 255 0.026368
20.32739067 39.18872070 -0.54711998 255 247 255 0.033936
20.32494926 39.18432999 -0.39478001 255 247 255 0.025391
20.32592964 39.18579102 -0.38159001 255 247 255 0.023194
20.33179092 39.19702148 -0.36743000 255 247 255 0.024414
20.32739067 39.18872070 -0.35376000 255 247 255 0.025391
20.32152939 39.17797852 -0.33960000 255 247 255 0.026368
20.32152939 39.17797852 -0.32593000 255 247 255 0.025391
20.32056046 39.17504883 -0.31176999 255 247 255 0.020020
20.36059952 39.25268936 -0.27124000 255 247 255 0.032959
20.36059952 39.25317001 -0.25757000 255 247 255 0.025391
20.37573051 39.28149033 -0.21558000 255 247 255 0.032959
20.38353920 39.29613876 -0.20190001 255 247 255 0.027588
20.41138077 39.34984970 -0.18822999 255 247 255 0.019043
22.94897079 44.24243164 -0.14917000 255 247 255 0.287907
22.95483017 44.25415039 -0.13403000 255 247 255 0.285466
22.95874023 44.26147079 -0.11792000 255 247 255 0.364584
22.96361923 44.27074814 -0.10278000 254 247 255 0.299626
22.96752930 44.27904892 -0.08716000 254 247 255 0.325261
22.97241020 44.28833008 -0.07153000 254 247 255 0.372641
22.97533989 44.29370117 -0.05591000 254 247 255 0.414649
22.98022079 44.30297852 -0.04028000 254 247 255 0.381674
22.98413086 44.31029892 -0.02466000 254 247 255 0.416358
22.98657036 44.31518936 -0.00903000 254 247 255 0.506950
22.98706055 44.31665039 0.00659000 254 247 255 0.590951
22.98804092 44.31859970 0.02222000 254 247 255 0.522332
22.98706055 44.31665039 0.03784000 254 247 255 0.487907
22.98657036 44.31518936 0.05347000 254 247 255 0.453223
22.98510933 44.31225967 0.06909000 254 247 255 0.395834
22.98461914 44.31127930 0.08472000 254 247 255 0.402197
22.98364067 44.30932999 0.10034000 254 247 255 0.383139
22.98413086 44.30981064 0.11597000 254 247 255 0.381674
22.98315048 44.30786133 0.13158999 254 247 255 0.365072
22.98168945 44.30541992 0.14673001 254 247 255 0.439307
22.98168945 44.30541992 0.16333000 254 247 255 0.422705
22.98119926 44.30443954 0.17847000 254 247 255 0.403418
22.98168945 44.30493164 0.19408999 254 247 255 0.408545
22.98364067 44.30884171 0.20972000 254 247 255 0.368978
22.98657036 44.31470108 0.22533999 254 247 255 0.420020
22.98950005 44.32006836 0.24097000 254 247 255 0.389731
22.99341011 44.32788086 0.25659001 254 247 255 0.379477
22.99731064 44.33520889 0.27221999 254 247 255 0.392172
23.00219917 44.34497070 0.28832999 254 247 255 0.376059
23.00757027 44.35522079 0.30396000 254 247 255 0.493523
23.01099014 44.36156845 0.31957999 254 247 255 0.484977
23.01586914 44.37134171 0.33521000 254 247 255 0.420996
23.01977921 44.37915039 0.35082999 254 247 255 0.449317
23.02466011 44.38843155 0.36646000 254 247 255 0.495720
23.02856064 44.39574814 0.38257000 254 247 255 0.525017
23.03149033 44.40161133 0.39771000 254 247 255 0.329412
27.61693954 53.24145889 0.49634001 254 247 255 0.137484
27.61693954 53.24097061 0.51489002 254 247 255 0.110369
27.61693954 53.24145889 0.53394002 254 247 255 0.108659
27.61352921 53.23413086 0.55249000 254 247 255 0.110369
27.61400986 53.23559952 0.57152998 254 247 255 0.112078
27.61304092 53.23413086 0.59008998 254 247 255 0.104997
27.61693954 53.24145889 0.60913002 254 247 255 0.143099
27.61841011 53.24390030 0.62720001 254 247 255 0.112322
27.61693954 53.24097061 0.64624000 254 247 255 0.104997
27.61400986 53.23559952 0.66478997 254 247 255 0.106706
27.61450005 53.23656845 0.68383998 254 247 255 0.119402
27.61352921 53.23461914 0.70239002 254 247 255 0.115740
27.61352921 53.23461914 0.72192001 254 247 255 0.132097
27.61400986 53.23608017 0.73999000 254 247 255 0.117693
27.61548042 53.23900986 0.75902998 254 247 255 0.126726
27.61499023 53.23754883 0.77758998 254 247 255 0.143099
27.61548042 53.23852921 0.79663002 254 247 255 0.172152
27.61548042 53.23804092 0.81519002 254 247 255 0.153841
27.61548042 53.23852921 0.83423001 254 247 255 0.166537
27.61646080 53.23999023 0.85277998 254 247 255 0.146761
27.61499023 53.23706055 0.87182999 254 247 255 0.173861
27.61548042 53.23900986 0.89038002 254 247 255 0.159213
27.61499023 53.23804092 0.90942001 254 247 255 0.153841
27.61400986 53.23559952 0.92749000 254 247 255 0.159213
27.61499023 53.23754883 0.94652998 254 247 255 0.143099
27.61548042 53.23852921 0.96557999 254 247 255 0.130388
27.61646080 53.23999023 0.98461998 254 247 255 0.162875
27.61499023 53.23804092 1.00268996 254 247 255 0.161166
27.61400986 53.23608017 1.02222002 254 247 255 0.139437
27.61548042 53.23900986 1.04077005 254 247 255 0.155550
27.61693954 53.24097061 1.05932999 254 247 255 0.162875
27.61157036 53.23120117 1.07788002 254 247 255 0.155550
27.61059952 53.22925186 1.09692001 254 247 255 0.130144
27.61548042 53.23852921 1.11597002 254 247 255 0.144808
27.61548042 53.23804092 1.13452005 254 247 255 0.155550
27.61499023 53.23804092 1.15307999 254 247 255 0.144808
27.61548042 53.23804092 1.17211998 254 247 255 0.157504
27.61693954 53.24145889 1.19115996 254 247 254 0.173861
27.61693954 53.24097061 1.20972002 254 247 254 0.150423
27.61450005 53.23656845 1.22827005 254 247 254 0.173861
27.61693954 53.24193954 1.24731004 254 247 254 0.184848

Excerpt of the original cloud index of one tree measurement. Each row represents one point, each number (from left to right) represents the XYX coordinate, RGB value and intensity respectively.