Width learning-based long-term target tracking algorithm

Abstract

The invention discloses a width learning-based long-term target tracking algorithm. The algorithm comprises the following steps of: establishing a width learning system; and carrying out tracking on the basis of the width learning system and establishing an acceleration steady feature algorithm-based whole-image detection mechanism. According to the algorithm, width learning system-based long-termtarget tracking is carried out, width learning architecture is relatively shallow in layer and relatively low requirement is put forward for calculation resources, so that deployment can be carried out on low-end equipment without losing great precision. According to the algorithm, an obtained target tracking model is high in training speed, small in reconstruction cost and low in time cost, andhas great superiority for detection of deformation, rotation and shielding in target tracking process. According to the algorithm, the SURF algorithm-based whole-image detection mechanism is applied,so that target information can be rapidly obtained when targets re-appear under the condition that the width learning system judges that the targets are lost, and the target positions can be updated so as to ensure that the tracking effect is more stable, robust and reliable.

Description

technical field [0001] The invention relates to target tracking in the technical field of computer vision, in particular to a long-term target tracking algorithm based on width learning. Background technique [0002] Target tracking has a very wide range of research and applications in many fields such as visual navigation, behavior recognition, intelligent transportation, environmental monitoring, battlefield reconnaissance, and military strikes. At present, the classic tracking method has poor adaptability to target scaling, rotation, occlusion, etc.; the more popular research is represented by the scale-invariant feature transformation method, namely the SIFT algorithm. The SIFT algorithm calculates the Gaussian filter of different windows at multiple scales. Image processing is used to achieve robustness to multi-scale scaling, rotation, blurring, etc. of the target, but it has a large amount of calculation and high complexity, and it is difficult to meet the real-time p...

AI Technical Summary

Problems solved by technology

At present, the classic tracking method has poor adaptability to target scaling, rotation, occlusion, etc.; the more popular research is represented by the scale-invariant feature transformation method, namely the SIFT algorithm. The SIFT algorithm calculates the Gaussian filter of different windows at multiple scales. Image processing is used to achieve robustness to multi-scale scaling, rotation, blurring, etc. of the target, but it has a large amount of calculation and high complexity, and it is difficult to meet the real-time processing requirements; while the mean shift theory uses the histogram as a feature and uses the kernel probability density to estimate The method to achieve target tracking, although it is robust to occlusion and rotation of the target, but it is not effective for large-angle rotation and hyperplane rotation; the target based on filter theory represented by particle filter and Kalman filter The tracking method takes filtering prediction as the core idea, and it is very robust to partial and complete occlusion of the target. However, there are still many problems in target tracking, such as large-scale scaling, rotation, hyperplane rotation, illumination changes, and partial occlusion of the target. , blurring, reacquisition after the target disappears from the field of view, etc.

In short, there is no good and complete solution at present. With the application of deep learning in the image field, the accuracy and real-time performance of target tracking have been greatly improved, and the shortcomings of general filtering algorithms have also been overcome. Improve

However, the deep learning network is complex, the training period is long, its construction and update process takes a long time, the amount of calculation is large, and the real-time tracking needs to be strengthened

Images