A novel depth measurement method for urban flooding based on surveillance video images and a floating ruler

In view of the low accuracy and high cost of existing urban flooding monitoring methods, this study proposed a method for measuring the depth of urban flooding using a combination of surveillance cameras and intentionally designed rulers. The method includes the binocular method and the subruler method, two implementation approaches based on a self-designed floating ruler to measure urban flooding depth. First, floating rulers were installed in the test area. Then, the pixel position of the floating rulers was obtained through surveillance cameras and the YOLOv5s object detection model. Finally, the water depth of urban flooding was obtained using two-pixel position to water depth transfer functions of the binocular method and the subruler method. The results showed that our proposed method precisely measured the depth of urban flooding in surveillance video frames, exhibiting an average relative measurement error of 8.541% and 5.250% and an average frame processing duration of 0.397 and 0.468 s. Compared with the existing method using machine vision and the binocular method, the subruler method can provide a low-cost and high-accuracy urban flooding monitoring solution, which has the potential for deployment in the field. The subruler method is recommended for use in areas prone to deep water accumulation.