Research on Target Detection Algorithm Based on NAO Robot

With the rapid development of humanoid robots in recent years, people are paying more and more attention to robot competitions. NAO robot, as a competition robot designated by the RoboCup standard platform group of the Robot World Cup, is deeply loved by scholars. NAO robots need to accurately detect and identify objects before they can track and grasp. Due to the limitations of related hardware, the detection results of NAO robots are greatly affected by the color difference between the object color and the background color, especially the light intensity of the environment. We study and compare the object detection algorithms based on NAO robots, and design a set of object detection algorithms based on NAO robots according to their hardware conditions. By designing a pre-processing scheme for images, the recognition of objects in images was improved, and the detection ability of NAO robots was improved.