River defense project slope collapse detection method based on deep learning

• Main Authors: MA RUOLONG, AN XINDAI, WU DI, YU LIXIN, XIE XIANGWEN, JIANG WENLONG, WANG YANYAO, SONG KEFENG
• Format: Patent
• Language: Chinese; English
• Published: 29.03.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a deep learning-based slope collapse detection method for a river defense project. The method comprises the following steps of: acquiring slope image data of the river defense project by using monitoring equipment; segmenting a slope region from the slope image data by adopting a lightweight U-net network model based on an attention mechanism; detecting a moving target in the slope area through a background modeling moving target detection algorithm; and finally, the moving target is judged through the lightweight convolutional neural network. According to the method, computer languages are embedded into intelligent monitoring camera equipment, the equipment is placed on a rod-shaped object on a river protection bank, slope data are automatically collected, slope collapse movement is automatically recognized, the problem of manual inspection and troubleshooting of the river protection project slope is solved, and the river protection project slope detection efficiency is improved. And