Logistics storage monitoring method based computer vision

Abstract

The invention is to use a computer vision method to design a target detecting and tracking algorithm, which improves target tracking precision and speed and is suitable for a requirement for intelligently monitoring the storage in a logistics park in real time. In order to improve accuracy and robustness of target tracking, the invention presents a nuclear circulation target tracking method based on an illumination invariant feature for avoiding the shortcoming that the nuclear circulation target tracking is easy to be influenced by illumination variation while only using pixel to compare nuclear-related similarity. The method comprises the steps as follows: calculating a partial sensitive column diagram to an input image firstly; extracting the illumination invariant feature, and then applying a nuclear matrix cycle structure to fast calculate a response confidence map of a target and a template in a frequency domain to obtain the accurate position of the tracked target. The algorithm presented by the invention has relative strong robustness while the illumination is drastically changed and the posture is adjusted.

Description

technical field [0001] The invention relates to an intelligent monitoring method based on computer vision, which belongs to the field of target detection and tracking in computer vision. By combining the illumination invariance feature of the video frame with the kernel method based on the cyclic matrix, the accuracy and real-time performance of target detection and tracking are improved, and it is suitable for intelligent security monitoring in public places, factories, warehouses and other places. Background technique [0002] Most of the video surveillance in logistics parks now adopts the traditional monitoring method. Cameras are installed in the monitoring area for 24-hour monitoring. The monitoring center performs video storage. The monitoring center is guarded by security personnel. , the security personnel handle the alarm event. [0003] Intelligent visual monitoring and related technologies are essentially different from traditional monitoring systems and technol...

AI Technical Summary

Problems solved by technology

[0014] The purpose of the present invention is to overcome the shortcomings of traditional monitoring and the problem of insufficient accuracy and speed in target tracking in existing intelligent monitoring, and propose a target tracking method that combines illumination invariant features and kernel cycle transformation. The proposed method It has strong robustness in the case of drastic changes in illumination and attitude adjustment

Images