Detection method of road area surveillance video blocked by leaves

Abstract

The invention belongs to the field of computer vision, and in particular relates to a method for detecting that a monitoring video of a road area is blocked by leaves. Firstly, a deep learning platform based on Mask R-CNN is established to train the network model of road area and leaf target, and then the detected road area results are optimized to meet the recognition expectations in urban video surveillance scenes; The display and understanding of the positional relationship of the three-dimensional space on the two-dimensional plane in monitoring, the defined occlusion rules, and finally the judgment result of whether there is a road area blocked by leaves. The present invention can replace the traditional manual method of browsing a large number of videos to check and confirm the occlusion of leaves one by one, get rid of the characteristic dependence based on subjective experience disputes such as the size of the leaf area, distribution area, distance, etc., and realize the three-dimensional space by simulating in two-dimensional space Occlusion-aware applications.

Description

technical field [0001] The invention belongs to the field of computer vision, and can be used in a detection system for monitoring the condition that a road area in a video picture is blocked by leaves. In particular, it relates to a detection method for road area monitoring video blocked by leaves. Background technique [0002] In the popularization and application of social public safety prevention video surveillance systems represented by "Safe City" and "Snow Bright Project", the construction of a large number of monitoring points within the urban area needs to be located on both sides of various public transportation roads. Due to the natural change of seasons, the improper use of some artificial cameras or the occurrence of accidents, there are leaves that cannot be ignored in the camera's monitoring field of view, resulting in the loss of monitoring target information on the road to varying degrees. Related business applications have caused obvious interference, and ...

AI Technical Summary

Problems solved by technology

[0003] The traditional leaf occlusion detection method generally uses manual feature extraction for leaf target recognition. Therefore, for engineers, the manual recognition ability of leaf target features poses a great challenge and is inefficient.

In terms of occlusion judgment, due to the complexity of the application of video surveillance scenes, whether the existence of leaves will cause visual interference will be affected by subjective factors. There is a lack of unified cognitive standards, and the application of prior knowledge is also limited. The overall effect is not ideal.

Images