Development of Manual Photo-Interpretation: Direct Observation (Landform and Land-Use Analysis) Using Interpretation Keys (Especially for GATE-Geospatial 2022)

Get unlimited access to the best preparation resource for CTET/Paper-1 : get questions, notes, tests, video lectures and more- for all subjects of CTET/Paper-1.

Examrace Books on Mapping, GIS, and Remote Sensing prepares you throughly for a wide range of practical applications.

Geographers have long been concerned with the use of aerial photographs as a major source of qualitative data, and applications of photo-interpretation techniques are highly varied. Based on the degree of sophistication with which the technique has been employed, one may distinguish two major directions of approach: (1) by direct observation; and (2) by indirect observation.

Both of these have been applied to extract qualitative data from aerial photographs for the natural environment as well as the artificial environment.

Direct Observation (Using Photo-Interpretation Key)

These include all applications using photo-interpretation to extract qualitative data of those phenomena which are directly observable on aerial photographs. Thus, parameters describing the external characteristics of the terrestrial environment are obtained. Such applications can be again sub-divided into two groups according to whether the natural or the artificial environment is studied, but in both cases the photo-interpreter concerned must focus on a particular aspect of interest (i.e.. with a specific objective in mind) and exercise his three levels of reference, i.e.. general, local and specific to arrive at an identification of features as accurately as possible. It should be stressed again that the use of aerial photographs does not dispense with field surreys completely but only minimises the need for them. Photo interpreted features need to be checked on the ground to establish the degree of accuracy. Hence, a usual approach of photo-interpretation is to develop a key for aiding the identification of objects in some sample areas and then applying it to other previously unsurveyed areas.

A photo-interpretation key has been developed on the principle of identification keys employed by field workers in the biological and physical sciences. The purpose is to help the observer to organise the information present in aerial photographs and guide him to correct identification of unknown objects. (The key is a systematic listing and description of the distinctive characteristics of the features to be identified supported by illustrations with annotations, which may present the vertical plan view in a stereopair or-as a single photograph and/or on oblique or even terrestrial views of the same features.

There are generally two types of photo-interpretation key depending on whether the principle of (1) selection or (2) elimination is employed for their construction.

Selection: In the first case, the different classes of phenomena are illustrated and described, and the photo-interpreter selects the one which comes closest to the features examined. In the second, a series of possible identifications are presented step by step, usually by means of a dichotomous system of paired choices, and the photo-interpreter is required to eliminate the one which is incorrect at each stage until he reaches the final answer. The usefulness of these keys has been the subject of argument among photo-interpreters, but it has been shown by a quantitative evaluation that if the material is reasonably well organised, no significant difference should exist between the two. However, one should note that the dichotomous key is designed to force the interpreter to work from the general to the by leading him through a mass of detail to a final answer. On the other hand, the selective key does not give great detail but treats the subject as a whole. It follows therefore that the dichotomous key is more difficult to prepare than the selective key because it has to provide rather detailed information at each stage. In forestry, dichotomous keys are usually employed. An excellent example is given by Sayn-Wittgenstein who made use of crown characteristics to identify tree species. An example of a good selective key is one dealing with industrial land use which was constructed by Chisnell and Cole. There are also other keys prepared for landform studies such as those by Powers and Waldo and Ireland.

There are also some more specialised types of key such as the regional keys and analogous area keys, both of which make use of the concept of regional geography. The former classifies the terrain into homogeneous units such as physiographic units whilst the latter attempts to develop the key in an accessible area for application to an inaccessible area, having been based on the assumption that every geographic region in the world has at least one analogous counterpart elsewhere. This is really an important approach which represents an early attempt to look for some general laws in regional geography.

In conclusion, one should note that the photo interpretation key only helps to identify, not to interpret; and the success of such keys must be judged, as Bigelow rightly points out, against the functions that they were originally intended for.

In the following discussion on the various past applications of the direct observation approach in photo-interpretation, two areas of interest to geographers can be distinguished:

  • landform analysis
  • land-use analysis

Landform Analysis

The external characteristics of the natural environment as imaged on black-and-white panchromatic film provide the impetus for a morphological approach which results in a description of the different landform types on a regional basis. Thus, photo-interpretation has been found to be particularly effective in providing information on the distribution of features associated with the following types of landform:

  • Volcanic Landforms such as cones, craters, crater lakes and lava flows;
  • karst landforms such as sinkholes, karst cones, ponors and poljes and the calculation of the ‘degree of karsting’ from drainage, hole-density and peat development in an area;
  • Glacial landforms such as eskers, kames, kettles, ice-dammed lakes, drumlins, moraines and glaciated valley forms;
  • landforms resulting from sub-aerial erosion and mass wasting such as gullies, landslides;
  • riverine landforms such as drainage patterns, forms of river channels, terraces and deltas;
  • Coastal Landforms such as tidal flats, marshes, beaches, sea cliffs, sea caves, arches, stacks, reefs, etc;
  • Aeolian landforms such as sand dunes and loess; and
  • Desert landforms u such as playas, bajadas, pediments, Hamada, and inselbergs.

All these landforms produced by aggradational as well as degradational forces can be easily interpreted from aerial photographs by evaluating the image characteristics of the various object՚s tonal contrasts and patterns. Based on these, individual landforms and features are then classified in accordance with the accepted terminology in geomorphology. The latter is a detailed landform classification based on field observations and therefore functions analogously as an interpretation key. As an example, it is not enough to identify a landform as a floodplain on the photograph; it is necessary to indicate at the same time whether it is a meander floodplain, a covered floodplain, a composite floodplain, or a bar meander flood plain, etc, according to the process of formation. Another example is the classification of mass wasting phenomena according to Sharpe՚s designation as slow flowage, rapid flowage, landslides, and subsidence. Such a distinction is usually very fine and is not always obvious from aerial photographs. But essentially the interpretation should take into account the processes, structures and stages of development of the landforms - an approach advocated by Lueder.

It is therefore clear that the interpreter has to exercise his good judgement based on a sound knowledge of geomorphology, geology, climatology and other related objects. Inevitably, features for one type of landform may be more easily interpreted than others; and between more contrasty environments, even the same image characteristic may differ in significance. Thus, Davis and heal stressed the differences between humid and arid regions. In the former, cohesive soils are usually associated with darker tones because of their higher moisture content, whilst in the latter, the clay surface of a dry playa often has greater reflectance than sand surfaces, and clay and salt have almost the same reflectance when dry. This fact necessitates some modifications of the photo-interpretation technique in analysing features of arid regions. Similarly, Verstappen has observed that ′ photo-interpretation becomes less revealing in areas where only rocks of similar resistance outcrop or where the effects of selective erosion are obliterated. He stressed the importance of a knowledge of climatic factors such as rainfall, temperature, humidity, etc, and drainage pattern as essential for the correct evaluation of the lithology and structure of the rock on which the land-forms are developed. The climatic factors, most notably rainfall, affect the intensity of dissection of the terrain and the kinds of weathering process (chemical or mechanical) prevalent in the area, as evidenced by a contrast in the characteristics of the landforms developed in the temperate humid, tropical humid, sub-tropical and sub-arctic arid regions. Obviously, the climatic factors cannot be obtained from aerial photographs alone and have to be supplemented elsewhere. On the other hand, the drainage patterns developed over an area are more readily apparent, from which one may infer the nature of the rock and the kind of geological control at work. As early as 1932, Zernitz classified drainage patterns into six basic types and twenty-four modifications. ″ A logical grouping of the more significant patterns, having taken into account their places of occurrence, is pictorially presented in below Figure, which displays the characteristics of the drainage pattern developed over lowland areas (anastomotic, Yazoo, dichotomic, braided and reticular) and highland areas; and in the latter whether the drainage pattern has been developed freely (dendritic, sub-dendritic, sub-parallel, parallel, and radial) or with structural controls (annular, trellis, angulate, rectangular, and contorted) . Of course, in actual situations, the drainage pattern developed over an area will not be so well defined and is likely to be a result of the mixture of two or more patterns. But this in conjunction with the grey tone criteria mentioned before will provide a strong basis for correct identification and evaluation of the images.

Major Types of Drainage Patterns (After Verstappen, …

On the other hand, a more detailed analysis of individual elements of the drainage pattern is possible using large-scale aerial photographs and involving the interpretation of river valleys and channels. It is possible to establish positively the type and size of the valley- and its main elements such as slopes, terraces, floodplains, etc, and the less directly recognisable characteristics such as the marshiness of the slopes and the valley floor, soils, etc. All these add up to give some understanding of the work of the water. Kudritskii et al have given very detailed instructions on the methods of interpreting the river valleys under different terrain conditions, such as in mountainous, hilly and flat areas; and in different cover types and seasons such as with a forest cover and in winter. 31 This can lead to the identification of the various types of valley such as rifts, gorges, canyons; V-shaped valleys; trough-shaped valleys; box-shaped valleys; trapezium-shaped valleys; unpronounced valleys; dry valleys and ravines. Similarly, the characteristics of the river channels such as meandering, forking, predominant types of channel formations, etc, can be interpreted from aerial photographs. It should be noted that only some of these features can be interpreted from aerial photographs by direct identification clues and that both the qualitative and quantitative aspects of the stereo model have been exploited for the extraction of useful information.

Land-Use Analysis

So far, only the landform aspect has been stressed, but, as has already been shown, the identification of a land-form type leads to inference on the rock types and structural geology of the area, or, in some cases, even on the soil, vegetation and land use. Each rock type has its geomorphological expressions, which means that a strong correlation exists between the two. Some general observations can be made as a guide to photo interpretation. Thus, landforms carved in igneous rocks exhibit the dominance of weathering along fractures, characterised by a coarse and regular dissection with a strong tendency for the development of a dendritic drainage pattern owing to the absence of any structural influences. Resistance to erosion differs according to the chemical composition of the rock types, as the basic rock has a high concentration of silicates, resulting in the basic rocks being less resistant than the acid ones. On the other hand, landforms carved in metamorphic rocks stand can be extracted from aerial photographs as a result of the control by scales:

  • the more generalised level of interpreting broad groups of land use from medium to small-scale aerial photographs (1: 10,000 or smaller) which results in a thematic land-use map
  • the highly detailed level of systematically determining the use of each parcel of land in the area under consideration from large-scale aerial photographs (e. g. 1: 1,000) which results in an inventory.

Each of these two types of approach meets different requirements and faces different problems during interpretation. Obviously, for the more detailed level of interpretation, greater accuracy is required, and as a result, the interpreter must be well versed in this technique and should possess highly specialised knowledge of the subject. There is also the need to stress the importance of the cultural background because land-use interpretation from one cultural situation cannot be directly transferred to another cultural situation without modifications. This is the same problem faced by geographers in transferring Western models of developed countries to non-Western developing countries; and photo-interpretation, to be accurate, must be based on an a priori model developed for the area in question.

For this detailed inventory approach, the accuracy of the land-use data interpreted depends heavily on the classification scheme used; and the success of the classification scheme, in turn, depends on the clarity of definition and description of each land-use category. It has been recognised that a single land-use classification system which will serve all users at all times is impossible to establish. A classification scheme designed for a land-use survey on the ground is very different in terms of detail from that for inventorying with large-scale aerial photographs, since in the latter the more ‘internal’ aspects of the land use will not be so readily visible. In addition, there is a weakness in the land-use classification scheme caused by confusion in the various concepts about land resulting in mixing what is to be measured and recorded. Land use has been variously viewed as the human activity on the land; the natural qualities of the land, land tenure, etc. geographical applications, the land-use classification can be morphological or functional or a mixture of both in design. The morphological scheme deals with the land cover rather than land use, as evidenced by the use of such terms as arable land, grassland, moorland, built-up area, etc - a good example being Stamp՚s scheme for the First Land Utilisation Survey of Britain in 1930, and the more detailed two-level scheme for the Second Land Utilisation Survey under Coleman in 1960. The functional scheme is activity-oriented, as evidenced by the use of such terms as agricultural, grazing, forestry, urban activities, etc. Anderson suggested that the latter may be more suitable as a general-purpose classification scheme for use with high-altitude photography or space imagery. However, for a more detailed survey of land use, it has been recommended by Clawson that a single or pure line concept should be used in a single classification scheme and an inductive approach be adopted. ″ This means that the interpreter should interpret land use in as much detail as possible for the smallest recognisable parcels of land so that the uses can be grouped into the categories most appropriate to his own investigation (i.e. to maintain a high degree of flexibility) .

As for the other more generalised level of land-use interpretation involving small-scale photographs, the problem of the classification scheme is equally applicable, although the demand for accuracy is less stringent. The number of categories of land use is obviously much fewer than in the detailed analysis case. The major source of error is due to grouping, resulting in some loss of detail caused by data aggregation within enumeration cells. A sampling design may also be employed to extract qualitative (as well as quantitative) land-use data along lines, at points or in areas on a random or stratified basis in order to produce small-scale land-use maps for large areas. Berry and Baker stressed the suitability of the stratified systematic unaligned sample design for extracting land-use data at points from aerial photographs, and a general small-scale land-use map can be constructed based on the data at these points. If this method is employed, sampling errors will occur.

These problems of extracting qualitative land-use data from aerial photographs at the detailed and generalised levels will be better appreciated if we examine some actual examples of applications. As it is easier to interpret crop types than it is to identify types of retail establishments, a division into rural land use and urban land use is justified in the following discussion.

Rural Land Use

Rural land use deals with such cover types as crops and vegetation in the natural environment and hence is concerned with such activities as agriculture and forestry. These cover types and activities are all directly recognisable from aerial photographs, and the photo-interpretation involved is relatively straightforward.

Crop Types and Farming Types

The identification of crop types in an agricultural area at different times of the year has special economic significance in a rural land-use survey. This can be carried out with large-scale (i.e. not smaller than 1: 10,000) aerial photographs. It has been pointed out that the farm crops can be differentiated on aerial photographs by the unique tonal and textural qualities of their photographic images and by objects which are commonly found in association with them. However, these tonal and textural characteristics of the crops tend to change at different growth periods, so that for the best results in the interpretation, aerial photographs should be taken three weeks before harvest. Although tonal values registered by each crop on aerial photographs are unique at certain intervals of growth, they do not only reflect the conditions inherent in the growth of crops but also combine the effects of ground moisture, sky conditions and other photographic characteristics. The use of a densitometer has been particularly favoured as a means of screening out the undesirable effects of standardising tone measurement. Texture qualities also tend to vary with the external characteristics of the crops and the manner in which they have been planted, cultivated and harvested. Textures can be fine or course, but a more detailed breakdown into ‘lined’ , ‘plaid-like’ , ‘Corduroy’ , ‘striped’ , ‘swath’ and ‘mottled’ textures was proposed by Goodman in the American context. ″ In Hong Kong, the differentiation between the paddy and the vegetables can be clearly made on textural characteristics alone (plate 16) . The use of the association of objects in identification is an indirect recognising characteristic which makes use of such relationships as the presence of straw stacks in the field indicating grain crops and the presence of haystacks indicating hay. Similarly, the presence of cattle or lanes leading to barns etc helps to establish the identity of rotation and permanent pastures. The cutting and tracking pattern in harvesting can also help to distinguish many crops or crop types.

Urban Land Use Urban

Land use reflects man՚s use of an artificial environment - the city where there is a concentration of people and a focus of human activities. Most of these activities; are undercover and hence are not directly visible from aerial photographs. Also, the diversity of these human activities causes greater difficulty in photo-interpretation than in the case of rural land use. But aerial photographs have been and still are favoured by urban geographers and town planners as a major source of urban information at varying levels of detail controlled by the photo-scale. Indeed, for more general uses, aerial photographs provide the necessary background data on the geographical setting of the urban environment. Thus, Witenstein incorporated this as a typical approach in the sequence of inventory, analysis and plan to be adopted by planners. He illustrated the usefulness of such an approach with an example from the United States - the town of Rockville, Maryland, where a planning problem to expand the facilities of the existing Central Business District to meet an unexpectedly large population overspilled from the nearby metropolitan area was in hand. His sequence of procedures involved first of all the examination of small-scale photo-mosaics to obtain an overall view of the urban environment; then stereopairs at large scale (by which he meant photographs at 1: 20,000 or larger) were used to study each part of the urban area intensively. Gross land-use types, such as residential, commercial, industrial, transport, vacant land, etc, were interpreted and mapped from the aerial photographs separately as a set of overlays on the photo-mosaics.

For each type, measurement was made of the size and capacity of use, thus obtaining statistical data (observed variables) on an aerial basis. Computation could then be made of the ratios of land used, land zoned, service availability and facility accessibility to residential, commercial and industrial needs, etc, thus yielding constructed variables for analysis and planning, with the two other procedures to follow in the sequence. Witenstein has carried out a ground check which supported the high accuracy of the photo land-use data and the usefulness of aerial photographs in developing detailed statistical data.

Developed by: