Method, system and equipment for marking key points of vehicle and medium

Abstract

The invention provides a method, system and device for marking vehicle key points and a medium. The method comprises the steps that the vehicle type of a to-be-marked vehicle is determined, a corresponding three-dimensional marking model is obtained according to the determined vehicle type, and three-dimensional key point marking is conducted on the to-be-marked vehicle; performing two-dimensional projection on the three-dimensional key point labeling result to obtain two-dimensional projection results of all key points; and performing two-dimensional key point labeling on the to-be-labeled vehicle, and determining a final key point labeling result of the to-be-labeled vehicle based on a two-dimensional key point labeling result and the two-dimensional projection result. According to the invention, the two-dimensional projection result and the two-dimensional key point labeling result are associated as the key point labeling result of the to-be-labeled vehicle, which means that the labeling result of the 2D labeling model and the two-dimensional projection result after projection of the 3D labeling model can be used as the key point labeling result of the to-be-labeled vehicle together; therefore, the problem that labeling errors exist when a 2D labeling model or a 3D labeling model is independently used can be solved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method, system, device and medium for marking vehicle key points. Background technique [0002] The detection and recognition of vehicles is one of the main tasks in the automatic video image recognition system, and it has a wide range of applications in security, vehicle ReID (Re-identification), video structuring and other applications. The labeling of vehicles is an important part of the recognition system. The accuracy of the labeling determines the accuracy of the algorithm. The information dimension and density of the labeling greatly affect the understanding of the target features by the deep neural network. [0003] Common vehicle annotations include 2D (2-dimensional) key point annotations, 3D (3-dimensional) annotations based on CAD (CAD-Computer Aided Design) models, etc. Among them, the 2D key point labeling can accurately obtain the position information ...

AI Technical Summary

Problems solved by technology

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a method, system, device and medium for marking vehicle key points, which is used to solve the difficulty and high cost of marking vehicles in the prior art. and prone to labeling errors

Images