Robot motion monitoring visual information fusion method based on MTLBP-Li-KAZE-R-RANSAC

Abstract

The invention relates to a robot motion monitoring visual information fusion method based on MTLBP-Li-KAZE-RANSAC, wherein the method is a distributed visual information fusion method for robot motionmonitoring in national defense military and civil fields, belongs to the field of target tracking, and is characterized by comprising the following steps: (1) determining a central element LBP value;(2) determining an 8-dimensional description vector in each sub-region; (3) calculating an approximate Euclidean distance between the two description vectors; (4) carrying out matching point pairs ofthe images, and determining a parameter matrix of projection transformation between the images; (5) carrying out dv operation on the remaining feature point pairs; (6) determining the number of sampling iterations; (7) determining a likelihood ratio; and (8) determining an optimal threshold. Compared with a traditional identification method, the method has the characteristics of high precision, high efficiency and low time delay, and the target identification speed is greatly improved. Accurate judgment is made for identification and elimination of wrong data, the confidence coefficient of accurate information is improved, and a more accurate identification result is obtained.

Description

technical field [0001] The invention relates to the field of target tracking, and mainly relates to a fusion method of distributed visual information for robot motion monitoring in the fields of national defense, military and civilian use. Background technique [0002] Target recognition and tracking has important application practices in the fields of unmanned driving, national defense and military affairs. For traditional target tracking, infrared or other machine vision methods are usually used to mark the key points or key parts of the target, and then mark the key points The detection, or the position of the mark in space or the transformation of its dynamic background, combined with the RANSAC algorithm to realize the tracking of the target object, but due to the change of the rotation angle or the light background, the state change of the target object from far to near. The change of its shape and size can easily lead to the loss or change of key points or key parts, ...

AI Technical Summary

Problems solved by technology

[0002] Target recognition and tracking has important application practices in the fields of unmanned driving, national defense and military affairs. For traditional target tracking, infrared or other machine vision methods are usually used to mark the key points or key parts corresponding to the target, and then mark the key points The detection, or the position of the mark in space or the transformation of its dynamic background, combined with the RANSAC algorithm to realize the tracking of the target object, but due to the change of the rotation angle or the light background, the state change of the target object from far to near. Changes in its shape and size can easily lead to the loss or change of key points or key parts, or when obtaining the data set of target information, it is easy to generate a large amount of abnormal data that needs to be eliminated, otherwise it is very easy to cause the failure of tracking the target

And because real-time data acquisition needs to sample a large number of data models, the processing speed of traditional systems cannot meet the real-time requirements, resulting in recognition delays, and the tracking of target objects is prone to failure

Images