Research and Application of Small Target Detection Algorithm Based on Bipedal Robot

With the rapid development and popularisation of robotics, bipedal robotics is widely used in various fields. The robot platform used in this paper is the NAO bipedal robot, focusing on the small target detection algorithms and applications of bipedal robots. Due to the limitations of the hardware of the bipedal robot, there are certain noise interference and blurring of the picture quality in the target detection, especially in the case of unstable external light intensity. In this paper, three target detection algorithms are compared and experimented to select the algorithm with excellent detection effect. And through the process of image vision pre-processing of the collected images further processing, analysis of the three colour space of the collected images, selected the least affected by the light of the YUV colour space as the NAO bipedal robot target detection of the colour space, and then the implementation of the image segmentation, image noise reduction process improves the ability of bipedal robot target detection.