Vehicle key point detection method based on gradient regression tree and apparatus thereof

Abstract

The invention discloses a vehicle key point detection method based on a gradient regression tree. The method comprises the following steps of acquiring an initial position of each key point of a picture to be detected, inputting the initial position of each key point into a trained classifier and detecting so as to acquire a final position of each key point; extracting a characteristic of each key point in each training sample; using the extracted characteristic to carry out classifier training. The invention also discloses an apparatus corresponding to the method. The method and the apparatus of the invention have an advantage that a gradient-regression-tree point-to-point regression strategy is used so that interferences of surrounding points are reduced.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a vehicle key point detection method and device based on a gradient regression tree. Background technique [0002] Most of the current intelligent transportation systems rely on license plate recognition to track and process traffic incidents, but the limitations of license plate recognition are too large, such as blurred license plates, license plate occlusion, or even fake or no license plates will cause license plate recognition to fail. If further Confirmation of vehicle logo, model, body color, vehicle annual inspection sticker features, and driver characteristics are of great help to vehicle tracking and identification. [0003] Vehicle key point detection can detect the 4 corner points of the car window and the key point position information of the car body. For example, the detection of the 4 corner points of the car window is helpful for the subsequent detection of vehicl...

AI Technical Summary

Problems solved by technology

[0004] In the prior art, there is a means of using a classifier to detect key points, such as Adaboost detection for the corner points of car windows or license plates. The classifier used in this feature detection method obtains the features in a certain area through a sliding window, and Comparing each feature to obtain the detection result is more effective for image processing methods that require block detection (such as license plate recognition), but if detection points are required, this method is easily interfered by other points on the edge. If there are other vehicles nearby, In the classification process, it is easy to detect that the position of the key point will shift to other directions, resulting in positioning errors

[0005] Therefore, the existing technology has the problem that the positioning is easily disturbed by the initial positioning position, and the key points are not easy to locate correctly.

Images