Vehicle tracking method based on unmanned aerial vehicle

Abstract

The invention provides a vehicle tracking method based on an unmanned aerial vehicle, and relates to the technical field of vehicle tracking. The method comprises the following steps: detecting vehicles in a video by using a deep learning mode, and extracting vehicle information in an image of a video frame; calculating to obtain the image position of the target vehicle, and calculating and predicting the area of the next frame of target vehicle; performing similarity comparison on a vehicle image obtained by detecting the region and a target vehicle, performing similarity comparison by using a Hamming distance by using a perceptual hash algorithm, and performing similarity comparison on a histogram color characteristic value and a local histogram characteristic value by using a Bhattacharyya distance formula; and finally, scoring the perceptual hash, histogram color feature and local histogram feature similarity contrast ratio by adopting a proper weight, and screening out the vehicle with the highest score as a tracking target vehicle. The method can reasonably and accurately track the vehicle in the video shot by the unmanned aerial vehicle, and has the advantages of cost reduction, high speed and high accuracy.

Description

technical field [0001] The invention relates to the technical field of vehicle tracking, in particular to a vehicle tracking method based on an unmanned aerial vehicle. Background technique [0002] The Yolo algorithm is a popular target detection algorithm in the visual field. Its innovation is to integrate the two stages of candidate area and object recognition, so that the structure of the algorithm is simple and the target detection speed is improved. The Yolov3 algorithm is the third version of the Yolo series. Its advantage adopts the idea of ​​regression. The input end directly puts the entire picture into the network, and the output end outputs the category and position of the regression preselection box. The terminal is realized by using a neural network. End-to-end object detection. The network structure of Yolov3 is based on the ResNet residual network structure. It uses multi-scale features for target detection, predictive classification and replaces Softmax los...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the deficiencies in the existing technology, the present invention proposes a vehicle tracking method based on UAV. Aiming at the problem that the existing vehicle tracking algorithm is not high in the tracking accuracy of the vehicle under the perspective of the UAV, the Yolov3 The algorithm uses vehicle detection as the feature extraction part of vehicle tracking, and integrates it into vehicle tracking to improve the accuracy of vehicle tracking

Images