A Hyperspectral Target Tracking Method Combining Spatial Spectral Features and Correlation Filtering

Abstract

The invention discloses a hyperspectral target tracking method combined with spatial spectrum features and correlation filtering. The method has good tracking effect when problems such as occlusion, similar objects and small tracking targets exist in the tracking. In order to achieve the above-mentioned purpose, the present invention includes the following steps: (1) first select the target search area through the initialization information of the first frame, extract the spectral features in the target search area, and train the spectral correlation filter Spe-CF; select according to the initial information The target area, and reconstruct the two-dimensional high-resolution target area according to the multi-band and multi-dimensional target area, extract the target HOG features in this area, and train the spatial correlation filter Spa-CF; (2) read the t-th frame image, Select the target search area according to the target position in frame t-1, and then estimate the target position range according to the spectral correlation filter Spe-CF trained in step (1). The Spa‑CF determines the final position of the target.

Description

technical field [0001] The invention relates to a hyperspectral target tracking method, in particular to a hyperspectral target tracking method combining spatial spectrum features and correlation filtering. Background technique [0002] Visual object tracking is one of the important research directions in the field of computer vision. It is a comprehensive technology that uses video sensors to estimate the position of the visual object in each frame of the image sequence by analyzing and understanding video image information to achieve object tracking. In recent years, researchers have introduced machine learning into target tracking, and regard target tracking as a binary classification problem. The classifier distinguishes the target from the background in the field of view, and the position with the highest confidence of the classifier is the target position. At present, the tracking method based on correlation filter has become a research hotspot in the field of target t...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a hyperspectral target tracking method that combines spatial spectrum features and correlation filtering for the deficiencies in the above-mentioned prior art. In this method, there are occlusions, similar objects, and small tracking targets, etc. Good tracking effect when there is a problem

Images