Photogrammetric analysis of streambed sediment size distributions: Adaptation and testing of a semi-automated workflow
Abstract
This study focuses on developing and applying a workflow to calculate sediment size distributions using a combination of photogrammetric approaches. We test and apply our workflow to quantify patch-scale dynamics in urban shale-gravel streambeds at West Creek, near Cleveland, Ohio. Our workflow inputs an image of a 0.5 m by 0.5 m patch and combines PebbleCounts, a previously published grain-sizing algorithm in the programming language Python, with manual measurements in ImageJ and additional calculations. Challenges with PebbleCounts include inaccuracies due to imprecise focal distances leading to incorrect pixel-tograin size scaling, measurement of sediment outside of the defined patch, leaves and litter, nondetection of some sediment particles within the patch, and the inclusion of partially sandy, vegetated, and subaqueous streambeds. We resolve these issues by cropping images to the defined patch, correcting the scaling post-PebbleCounts processing, and using ImageJ to measure uncounted grains manually. For five test images, PebbleCounts alone produces a d50 value of 9.28 mm and a d90 value of 19.18 mm with 1,302 pebbles counted, whereas our developed workflow produces a d50 value of 6.89 mm and a d90 value of 16.54 mm with 1,498 pebbles counted. The combination of PebbleCounts and ImageJ in our workflow allows for comprehensive documentation of gravel and cobble size distributions, facilitating analysis of spatial variability and temporal patch dynamics with greater accuracy and robustness. The adaptation of PebbleCounts beyond its original design demonstrates the flexibility and potential of photogrammetric grain size algorithms for diverse applications. Through the development of a customized workflow, we have achieved results that enhance the efficiency of pebble counting and measurement compared to fully manual field or photogrammetric methods. This optimization enhances the capability to document and analyze sediment size distribution using photogrammetric and computational techniques, enabling better assessments of geomorphological changes in urban streams.
Primary Faculty Mentor Name
Anne Jefferson
Status
Undergraduate
Student College
Rubenstein School of Environmental and Natural Resources
Program/Major
Environmental Sciences
Primary Research Category
Life Sciences
Photogrammetric analysis of streambed sediment size distributions: Adaptation and testing of a semi-automated workflow
This study focuses on developing and applying a workflow to calculate sediment size distributions using a combination of photogrammetric approaches. We test and apply our workflow to quantify patch-scale dynamics in urban shale-gravel streambeds at West Creek, near Cleveland, Ohio. Our workflow inputs an image of a 0.5 m by 0.5 m patch and combines PebbleCounts, a previously published grain-sizing algorithm in the programming language Python, with manual measurements in ImageJ and additional calculations. Challenges with PebbleCounts include inaccuracies due to imprecise focal distances leading to incorrect pixel-tograin size scaling, measurement of sediment outside of the defined patch, leaves and litter, nondetection of some sediment particles within the patch, and the inclusion of partially sandy, vegetated, and subaqueous streambeds. We resolve these issues by cropping images to the defined patch, correcting the scaling post-PebbleCounts processing, and using ImageJ to measure uncounted grains manually. For five test images, PebbleCounts alone produces a d50 value of 9.28 mm and a d90 value of 19.18 mm with 1,302 pebbles counted, whereas our developed workflow produces a d50 value of 6.89 mm and a d90 value of 16.54 mm with 1,498 pebbles counted. The combination of PebbleCounts and ImageJ in our workflow allows for comprehensive documentation of gravel and cobble size distributions, facilitating analysis of spatial variability and temporal patch dynamics with greater accuracy and robustness. The adaptation of PebbleCounts beyond its original design demonstrates the flexibility and potential of photogrammetric grain size algorithms for diverse applications. Through the development of a customized workflow, we have achieved results that enhance the efficiency of pebble counting and measurement compared to fully manual field or photogrammetric methods. This optimization enhances the capability to document and analyze sediment size distribution using photogrammetric and computational techniques, enabling better assessments of geomorphological changes in urban streams.
