A Novel Python Video Processing Program for Identifying and Quantification of Soil Cracks in Real Time

The presence of soil cracks can lead to higher rainfall water infiltration, which will in turn cause instability in ground engineering infrastructure. Maintenance of ground infrastructure is hence essential, and such cracks are often identified through manual observations, that are laborious and not easily quantifiable. One of the ways could be to develop an automated program that can capture cracks through video processing. The objective of this study is to develop a simple video processing technique for soil crack sorting and its quantification in real-time. The new program automates the sorting of a region of cracks according to the given boundaries and quantifies the number of cracks in that region. A stepwise strategy is demonstrated to efficiently sort cracks and compute the crack intensity factor of real-time video with a negligible delay. Python script takes a frame from a video and analyzes it. It first automatically extracts the required portion of a video and computes cracks in this region. It also identifies the frames which have cracks and gives corresponding time and location to the administrator. Such a program can be useful in the future for analyzing videos obtained from UAV monitoring of the large area and, thus, identify vulnerable areas for maintenance.