A vehicle parking violation detection method based on object detection and semantic segmentation

Abstract

The invention discloses a sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation, which belongs to the field of deep learning and computer vision. In a training stage, a semantic segmentation network and a target detection network are respectively trained to obtain a semantic segmentation model and a target detection model; In the testing stage, the mixed Gaussian background model is used to extract the urban road background map, and then the semantic segmentation algorithm and the semantic segmentation model are used to perform semantic segmentation to obtain the semantic segmentation map. Then, the perceptual hash algorithm is used to determine whether the camera is rotating, and then the target detection is used. The algorithm and the target detection model perform vehicle detection, mark the vehicle detection frame, and finally compare the vehicle detection frame with the semantic segmentation map. The invention solves the problem that the existing vehicle illegal parking detection method has a high misjudgment rate, poor real-time performance, needs to manually demarcate the illegal parking area on the sidewalk, and when the camera rotates, it needs to manually demarcate the illegal parking area, which is not suitable for rotating the camera. The problem.

Description

technical field [0001] The invention belongs to the fields of deep learning and computer vision, and relates to a method for detecting illegal parking of sidewalk vehicles based on target detection and semantic segmentation. Background technique [0002] With the rapid development of economy and urbanization, the total number of roads and vehicles in various cities in my country is increasing, and the illegal parking behavior of vehicles is also increasing. The illegal detection of vehicles in urban road surveillance images or videos has become an important part of urban management. an important task. Although high-definition surveillance cameras have been deployed at most intersections, the amount of video generated every day is increasing. It is time-consuming and laborious to perform real-time video monitoring or offline processing manually, and is prone to delays and omissions. Therefore, there is an urgent need to find an efficient way to The method assists manual monit...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation, which solves the problem of high misjudgment rate and poor real-time performance in the existing vehicle illegal parking detection method, which requires manual calibration of sidewalk illegal parking. Parking area, and when the camera rotates, it is necessary to manually re-calibrate the illegal parking area, which is not applicable to the problem of rotating the camera

Images