Intelligent identification, positioning and tracking method and system based on deep learning

Abstract

The invention provides an intelligent recognition and positioning tracking method based on deep learning. The method comprises the steps of establishing a camera imaging geometric model through calibration of a binocular camera, correcting lens distortion, and obtaining an image through the corrected binocular camera; inputting the obtained image into a trained improved YOLO-Lite network for target detection, determining a target frame, and realizing target identification and positioning; converting the image of the target frame from an RGB space to an HSV space, and calculating a color probability distribution diagram; and selecting an initialized search window, calculating the centroid position of the search window, adjusting the size of the search window according to the centroid position until an iteration termination condition is met, taking the size and the center position of the search window of the previous frame as the initial search window of the image of the next frame, and adaptively adjusting the size of the window to realize target tracking. According to the method provided by the invention, accurate target identification and positioning can be realized in the presence of factors such as a complex background and illumination variation, effective tracking is carried out, and real-time performance is ensured.

Description

technical field [0001] The field of unmanned aerial vehicles of the present invention particularly refers to a method and system for intelligent identification, positioning and tracking based on deep learning. Background technique [0002] In recent years, small UAVs have developed rapidly in the direction of autonomy and intelligence, and their application range has become wider and wider, especially in the absence of satellite navigation. The demand for use is increasingly strong. Target monitoring is one of the important applications of UAVs. Using stereo vision methods can realize the target tracking and positioning of UAVs in complex environments. By building a binocular vision platform and using an appropriate target tracking algorithm, the distance and position of the target, the direction, trajectory, and size of the target can be effectively obtained. But at present, for target tracking, it is not possible to grasp the characteristics of the target in the presence ...

AI Technical Summary

Problems solved by technology

[0003] The main purpose of the present invention is to overcome the above-mentioned defects in the prior art, and propose a method for intelligent identification and location tracking based on deep learning, which can be used in the presence of factors such as complex backgrounds, illumination changes, similar color objects, and occluders. Accurate target recognition and positioning, effective tracking, and real-time performance

Images