Vision-based underwater cluster neighbor tracking measurement method

Abstract

The invention provides a vision-based underwater cluster neighbor measurement method. Based on a Meanshift tracking algorithm, a target color probability density distribution diagram is adopted as a feature of target search, and the algorithm is converged to a target real position through continuous iteration, so that a target tracking function is realized; underwater neighbors are positioned by using similar triangular target ranging and tangent function target angle measurement. The algorithm is small in calculation amount, good in real-time performance and insensitive to external environment interference.

Description

technical field

[0001] The invention is a vision-based underwater cluster neighbor tracking measurement method, which is mainly applied to underwater autonomous vehicles with limited visual information. Background technique

[0002] In recent years, autonomous underwater vehicles (AUVs) have attracted much attention with the implementation of the national marine strategy, and AUV clusters are a basic and innovative model of the deep integration of "artificial intelligence + marine unmanned systems". technology has attracted the attention of many researchers. Make multiple AUVs cooperate with each other like wolves, bee swarms, and fish swarms in nature to complete complex group movements such as foraging, nesting, enemy defense, and migration. Using independent individuals with very limited perception and action capabilities, by following some simple Behavioral rules and joint completion of complex tasks are the core issues in the study of cooperative control of underwater ...

Images