Method and device for detecting vehicle

Abstract

The invention discloses a method and device for detecting a vehicle, and relates to the field of computer vision. The specific implementation scheme comprises the steps of acquiring a to-be-detected image; inputting the to-be-detected image into a pre-established vehicle detection model to obtain a vehicle detection result, the vehicle detection result including category information, coordinate information, coordinate credibility and coordinate error information of a detection frame, and the vehicle detection model being used for representing a corresponding relationship between the image andthe vehicle detection result; according to the coordinate credibility of the detection frame, selecting the detection frame from the vehicle detection result as a to-be-processed detection frame; andbased on the coordinate information and the coordinate error information of the to-be-processed detection frame, generating the coordinate information of the processed detection frame. According to the embodiment, the accuracy of the coordinate information of the detection frame is improved, and detection errors caused by inaccurate detection of the vehicle detection model are reduced.

Description

technical field [0001] Embodiments of the present disclosure relate to the field of computer technology, and in particular to computer vision technology. Background technique [0002] In recent years, with the rapid growth of the number of traffic vehicles, traffic supervision is facing great challenges. Vehicle object detection, as a key technology for constructing video surveillance of traffic conditions, has received extensive attention from researchers at home and abroad. Usually, when monitoring the actual road scene, it will encounter the situation that the vehicle is too long and the detection is inaccurate. For example, for trailers on urban roads, the detection model usually used to detect vehicles uses a limited number of trailer images in the training set during training, which will cause the detection of the detection model to appear that the detection frame only frames the front of the vehicle or only frames the vehicle The case of the rear hanging part. The ...

AI Technical Summary

Problems solved by technology

Usually, when monitoring the actual road scene, you will encounter the situation that the vehicle is too long and the detection is inaccurate

For example, for trailers on urban roads, the detection model usually used to detect vehicles uses a limited number of trailer images in the training set during training, which will cause the detection of the detection model to appear that the detection frame only frames the front of the vehicle or only frames the vehicle The case of the rear hanging part

Images