Environment situation assessment method for autonomous grabbing of underwater visual target

Abstract

The invention relates to the technical field of computer vision, and particularly discloses an environment situation assessment method for autonomous grabbing of an underwater visual target, which aims at underwater mechanical arm operation and comprises the following steps of: calibrating dangerous object position information and danger coefficient assessment grade information in an underwater environment in advance to train a target detection and identification network N1; and enabling the trained target detection and recognition network N1 to recognize the position of a dangerous object inan underwater source image shot by any monocular camera and the danger coefficient evaluation level of the dangerous object, and generating a corresponding environment situation evaluation graph in combination with the depth estimation image, so as to form loss with an environment situation evaluation truth value image, and therefore, the information fusion network N2 is optimized, and an environment situation assessment graph generated by the optimized information fusion network N2 can be used as an important support for subsequent underwater environment operation tasks such as path planningand autonomous obstacle avoidance grabbing, so that the robot can be guided to realize the optimal behavior at a higher level.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to an environmental situation assessment method for autonomous grasping of underwater visual targets. Background technique [0002] In order for the robotic arm to perform tasks in a complex and dynamic underwater environment, it must have the ability to intelligently analyze the external environment. Risk assessment of the surrounding environment is an important guarantee for the safety and completion rate of manipulators and underwater vehicles. At the same time, environmental situation assessment can also be regarded as the process of the robot understanding the surrounding environment and modeling the environment, which is an important basis for subsequent path planning. Compared with traditional artificial potential field methods, behavior decomposition methods, and optimization-based algorithms, environmental situation assessment can guide robots to achieve optimal be...

AI Technical Summary

Problems solved by technology

[0003] The present invention provides an environmental situation assessment method for autonomous grasping of underwater visual targets, and the technical problem to be solved is that currently, for the underwater manipulator operation, the factors considered are not comprehensive enough, and there is no relatively complete environmental situation assessment strategy , which can reasonably plan the path for the robot and realize autonomous obstacle avoidance and grasping

Images