SURFACE RECONSTRUCTION AND CHANGE DETECTION 
    
FOR AGRICULTURAL PURPOSES BY CLOSE RANGE PHOTOGRAMMETRY 
Katharina Helming^, Wolfgang Jeschke', Juergen Storl 
Technical University of Berlin, EB 9 
Germany 
Commission V 
^now at Zentrum f. Agrarlandschafts- und Landnutzungsforschung, Müncheberg 
1 
ABSTRACT: 
now with Siemens Nixdorf Information Systems (SNI), Munich 
From 1987 through 1991 a so-called Interdisciplinary Research Project had been carried out at the TU Berlin. This 
project brought together scientists from various disciplines, i.e. photogrammetrists, soil scientists, geologists, and 
material research scientists. Their common aim was to develop working solutions for registration, numerical 
evaluation, and interpretation of surfaces from close-range photogrammetry and Raster Electron Microscopy 
(REM). The photogrammetric task within this project was to provide Digital Elevation Models (DEMs) of a 
representative selection of typical surfaces that were to be evaluated for their information content by the project 
partners. This paper reports on analog (and digital) data acquisition, photogrammetric processing, DEM determina- 
tion and interpretation, and conclusions of the joint work of photogrammetric and water-agricultural scientists. 
The microrelief of soil before, during, and after more or less heavy rain had to be defined by photogrammetric 
methods. Knowing this relief, the amounts of rain, and the related infiltration and draining rates it should be 
possible to give recommendations for an erosion minimizing treatment of soil. 
KEYWORDS: Close-range, DEM, Image Matching, Soil Erosion, Stereoscopic, Surface Measurement 
INTRODUCTION 
Soil erosion problem 
Soil erosion due to water impact is an agricultural and 
- increasingly - an ecological problem since fertile soil 
is washed away from the fields and, likewise, nutrients 
and pesticides are flushed to the outflows. In order to 
fight the causes of erosion a detailed knowledge of the 
erosion process - and its releasing factors as well — is 
essential. Erosion most frequently starts on bare soil 
where rain energy leads to a silt of the upper soil layer, 
thus hindering infiltration and supporting runoff in- 
stead. The so-called splash process that is induced by 
the raindrop impact force results in a sealed surface. 
Any raindrop hitting the ground causes development 
of a crater on the soil surface where the vertical forces 
of the drop are transformed into radial forces. When 
reaching maximum height, the crest of a crater ejects 
little droplets consisting of water and small soil particles 
(FERREIRA & SiNGER 1985). These particles are able to tamp 
the draining pores and to form with their layer-like 
deposit an infiltration-hampering seal. 
This interrill erosion is a rain energy driven process which 
is strongly influenced by surface conditions and struct- 
ures. Leaving out the plants covering the ground it is 
the microrelief of the surface that plays the most impor- 
tant role in this process since it has a decisive influence 
on the formation of the surface. Microrelief is the defini- 
tion for the relief range from 2 mm up to 200 mm 
that is formed by secondary tillage. It is non-directional 
and is characterized by aggregates and clods (ROMKENs 
& Wang 1984). Much work has been done to analyze 
its important influence on erosion and to prove that 
an increase in surface roughness will enhance infiltration 
and lead to a decrease in runoff (Jonson ET AL. 1979, 
STEICHEN 1984). Possible reasons are: 
the rougher a microrelief, the lower is the frequency 
of raindrop impact with reference to surface area 
and time as well as its effective angle. This leads to a 
decline of the sealing process (LINDEN ET AL. 1988) - 
i.e. this process is not controlled by the kinetic energy 
of the rainfall. It is just its effective (i.e. normal) 
component that causes interrill erosion. 
* a rougher microrelief leads to a higher depressional 
Storage, thus delaying the runoff begin (Moore & 
LARSON 1979). 
Measurement efforts 
Up to now no exact quantitative evaluation has been 
possible, since scientists were lacking methods for the 
precise measurement of the microrelief (elevation mea- 
sures for the surface are needed that are collected with 
a spacing of less than 2 mm and have a height precision 
of 0.2 mm or better). Moreover, the measurements have 
to be restricted to contact-free methods in order to 
permit multitemporal surveys (e.g. linked to artificial 
rain tests) of a surface. 
There have been some relief measurements over the 
last 40 years. First generation mechanical relief meters 
consisted of a board with moving steel needles (BurweLL 
ET AL. 1963) which vertical position could be measured. 
Spacing used was from 50 mm to 100 mm. Further 
development of this technique resulted in automation 
(i.e. photographic or electronic registration) and a spa- 
cing scaled down to 5 mm (Currence & Lovery 1970, 
Moore & Larson 1979, Tessier ET AL. 1989). This spacing 
and, more important, the vertical resolution of more 
than 1 mm still hindered detailed evaluations and, in