|NOUWAKPO, SAYJRO - Purdue University
|Huang, Chi Hua
Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/26/2011
Publication Date: 2/1/2012
Citation: Nouwakpo, S.K., Huang, C. 2012. A simplified close range photogrammetric technique for soil erosion assessment. Soil Science Society of America Journal. 76:70-84. DOI:10.2136/sssaj2011.0148.
Interpretive Summary: Digital photogrammetry is the remote measurement of geometry from digital photographs. The availability of low cost digital camera makes surface reconstruction using digital photogrammetry a great tool for soil erosion assessment. The technology requires the user to accurately survey control points or markers that are used to provide scale and accuracy to the digital photogrammetric system. Currently, survey grade equipments are used to measure these markers. The ideal tool for a soil scientist is one that is cheap, light weight and can fit in his backpack so that he can monitor channel development, elevation changes, etc as often as needed. In this paper we propose a two-step photogrammetric technique for soil erosion assessment that does not need any survey equipment. The method uses a modified photogrammetric procedure to derive the coordinates of positional markers first and then followed by a conventional photogrammetric surface reconstruction procedure. We found that the control points obtained using the photogrammetric procedure were as accurate as the one surveyed using a total station. We have tested our methodology in the lab and used it to digitize ephemeral gullies in two fields in Kansas. We also assess the accuracy and sensitivity of digital photogrammetry to detect soil erosion and found that digital photogrammetry was ideal for tracking drastic changes in elevation such as the ones observed in gullies and rills. This simplified photogrammetry procedure makes it feasible to make accurate surveys of soil erosion features without the need to use a traditional surveying instrument. We are currently using this tool in the laboratory to routinely track channel network development.
Technical Abstract: Surface reconstruction using digital photogrammetry offers a great advantage for soil erosion research. The technology can be cumbersome for field application as it relies on the accurate measurement of control points often using a survey grade instruments. Also, even though digital photogrammetry has been used in many soil erosion researches, its sensitivity in detecting soil elevation changes has rarely been assessed. This paper aims at simplifying the digital photogrammetric procedure for soil erosion research and assessing the sensitivity of this technology to detect soil surface elevation changes. To simplify the technology, we proposed a two-step photogrammetric technique for soil erosion assessment. The method combines a photogrammetric procedure for control points generation in a first step followed by a conventional photogrammetric Digital Elevation Model (DEM) extraction procedure. The performance of the method was assessed in the laboratory and tested in the field to digitize ephemeral gullies. In the accuracy test of the first step of our methodology, we found that the maximum Length Measurement Error (LME) was 3.4 mm while the maximum angular deviation from the vertical or horizontal axes was 0.93°. The maximum error between control points coordinates generated by photogrammetry and those generated by a survey grade total station was 0.026 m on the horizontal axes and 0.01 m on the vertical axis. We also found that the sensitivity of digital photogrammetry in detecting soil surface elevation changes was similar to that of a laser scanner when the detection was performed on smooth soil surfaces and when the standard deviation of the elevation changes was approximately three times the precision of the photogrammetric DEM. This new development reduces the need to have a regular survey instrument in order to obtain the coordinates of control points in conventional photogrammetry procedure.