The Research Of Method To Test Robot Vision And Software Tools Which For Aiding To Test The Ability Of Robot Vision

Robots can replace humans in a variety of dangerous, heavy and boring work, and it has broad application prospects in natural disaster relief and nuclear leaks, and alien polar expeditions, military reconnaissance and combat, logistics and industrial manufacturing automation, intelligent civilian vehicles helping old disabled, etc. Currently the world is developing various robots. The perception of robot in an unknown environment is one of the key decisions robot autonomy technique, and robot visual ability is an important part of perception. At present domestic and foreign researchers are working on how to improve the cognitive abilities of the robot, and let robot adapt its working environment better and intelligently. Researchers has carried out extensive research in robotics-aware technology, and developed many theoretical methods and techniques. However, how to evaluate the pros and cons of various methods and techniques is a question which puzzled many researchers. Researchers often use only a limited set of database in their own environment to prove their methods and techniques better than others. This evaluation approach has limitations, and perception of individual robots can’t be relatively unified in accordance with the same standard one. People develop many methods and techniques to improve the visual ability of robot, but robots still cannot adapt to the reality of work better.This paper is supported by the project of Robot Environmental Awareness Test System. We research methods to test robot vision ability and design software tools which for aiding to test the ability of robot vision. This software can generated test images of different effects controllably, quantifiably and randomly. So this software provide a reference for integrated assessment of the robot vision ability. Good robot vision system is conducive to enhancing the ability to test the robot’s autonomy. We hope this software we designed can improve the autonomy of robots, and promote the process of machine vision applications in engineering.The main contents include, the research of method to test robot vision and software tools which for aiding to test the ability of robot vision and design experiment of robot vision ability test. We combined the visual ability test methods with software we have proposed in this paper and design experiment for testing robot vision ability. In the experiment, we chose an unmanned vehicle for the test object and traffic signs for testing content to establish a database of virtual verification test methods and evaluation system with the actual environmental test combination.