Long-time unmanned aerial vehicle tracking and positioning method, system and device based on stereoscopic vision

Abstract

The invention belongs to the technical field of visual tracking, particularly relates to a long-time unmanned aerial vehicle tracking and positioning method, system and device based on stereoscopic vision, and aims at solving the problems that an existing visual tracking method is poor in target tracking and positioning real-time performance and low in precision. The method of the system comprisesthe following steps: acquiring a binocular vision image pair at a moment t; preprocessing the image pair, and matching two images in the preprocessed image pair through a binocular stereo matching algorithm to obtain three-dimensional point cloud data; clustering the three-dimensional point cloud data to obtain a candidate region of the to-be-positioned target; for each candidate region, obtaining a first matching degree and a second matching degree through a template matching algorithm and a preset first method; and if the first matching degree of each candidate region is greater than a setfirst threshold and the second matching degree is less than a set second threshold, taking the candidate region as a target region of the to-be-positioned target. According to the invention, the real-time performance and precision of target tracking and positioning are improved.

Description

technical field [0001] The invention belongs to the technical field of visual tracking, and in particular relates to a long-term UAV tracking and positioning method, system and device based on stereo vision. Background technique [0002] In recent years, discriminative learning methods have been widely used in visual tracking research. Visual tracking is a key component of many computer vision systems and can be naturally formulated as an online learning problem, that is, given an initial image patch containing an object, train a classifier to distinguish the object from its surroundings. The classifier can be fine-tuned at many locations to detect it in subsequent frames, with each new detection providing a new image patch to update the model. The current mainstream visual tracking methods include tracking methods based on correlation filtering and tracking methods based on neural networks. [0003] The tracking method based on correlation filtering is to train a correlat...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem that the real-time performance of target tracking and positioning is relatively low due to the large amount of calculation, the error-prone template update during the long-term tracking process, and the large demand for training samples in the existing visual tracking methods. Poor, low precision problems, the first aspect of the present invention, a long-term UAV tracking and positioning method based on stereo vision is proposed, the method includes:

Images