Target tracking method based on learning and speeded-up robust features (SURFs)

Abstract

The invention discloses a target tracking method based on learning and speeded-up robust features (SURFs), and mainly solves the problem of target tracking failure due to quick change or occlusion of a target in the prior art. The target tracking method is implemented by the following steps: (1) inputting a first frame of a video, manually marking the target to be tracked, and taking the marked target as a target template; (2) tracking the target through a tracking-online learning-detection model; (3) judging a target tracking and detection result; (4) extracting the target template and the SURFs of the current frame of the video; (5) matching the obtained SURFs by utilizing the Euclidean distance; (6) outputting a target tracking result, and updating the target template; and (7) circularly performing the steps (2) to (6) until the video is ended. Compared with the prior art, the target tracking robustness is improved under quick change or occlusion of the target.

Description

technical field [0001] The invention belongs to the technical field of image processing, relates to a video target tracking method, and can be applied to intelligent monitoring, target tracking and man-machine interface. Background technique [0002] Target tracking of sequence images is an important part of the application of image processing technology. It refers to analyzing the input video image sequence to determine the position of the target in each frame and obtain relevant parameters. Target tracking is one of the key technologies in computer vision. It integrates image processing, pattern recognition and artificial intelligence. It is widely used in many aspects such as robot visual navigation, safety monitoring, traffic control, video compression and weather analysis. For example, military aspects have been successfully applied to imaging guidance of weapons, military reconnaissance and surveillance, etc.; civilian aspects, such as visual surveillance, have been wi...

AI Technical Summary

Problems solved by technology

In this method, the template image and the pixels of the area to be matched are arranged in a circle into multiple sub-windows, and the circular template matching criterion is used to ensure that the target has translation and rotation invariance, and the Kirsch operator is used to calculate the edge strength of each pixel in the template and the tracking window Value, the sum of the grayscale matching value and the intensity matching value is used as the matching result, and the position of the best matching value is determined as the position of the tracking target, but its defect is: when the target is occluded, the matching error leads to tracking failure

The tracking method includes: by default, the zeroth frame search window is as large as the image, the first frame image is recognized and the bounding box is obtained, and then the search window is predicted. It uses the image processing method to calculate the bounding box and its feature points, and at the same time On the basis of target tracking, a method of predictable search window is proposed, and the movement prediction and tracking of the marked target are carried out, which narrows the search range. Although this tracking method has a certain effect on improving real-time performance, when the moving target occurs When occlusions or rapid changes occur, accurate tracking cannot be achieved using the above predictive search window method

Images