Monitoring of soilroughness caused by rainfall using stereo-photogrammetry

ABSTRACT: In this study we tested stereo-photogrammetry under field conditions to detect changes of soil surface roughness caused by rainfall upon two agricultural soils surfaces - a seedbed plot and a harrowed plot. Three rainstormswith intensities of 60, 80 and 80 mm/h, respectively,were simulated on two plots of 1m2area and DEMs were generated before and after each rainfall. Using, DEMs, Three groups of roughness indices were calculated for each situation: (1) random roughness (RR), reduced random roughness (RR_rd) and corrected random roughness (RR_cr), (2) Tortuosity (Tb) and (3) limiting elevation difference (LD) as well as limiting slope (LS). In addition, surface variograms were calculated to characterize the microtopography at the beginning and end of the rainfall experiment.A detailed analysis of the factors influencing the stereo-photogrammetry process and the calibration of the system showed the importance of a stable support bearing the two cameras as well as a calm weather condition. Moreover, the effect of the correlation window size to optimize the DEM quality should be considered. All investigated roughness indices decreased as a result of rainfall on both plots; especially,LD and RR_cr were negatively correlated to cumulative rainfall.However, due to some difficulties in estimating Tb, no statement on its usability for detection of roughness changes is made in this study. The analysis of surface variograms showed that the heterogeneity of soil surface can be decreased by rainfall, while spatial correlation can be increased.If some basic provisions are considered to guarantee the image quality, stereo-photogrammetry is a promising method to detect soil surface roughness and its changes caused by rainfall.