| Vision sensor can obtain much more environment information than other sensors. So vision sensor may provide reliable information ensurence for robotic control and decision-making. Nowadays, binocular vision sersor is widely applied. Because depth information of objects and other 3D information can be obtained by binocular vision sensor. Stereo vision is one of international hot research field in nowadays. The Pioneer 3-DX mobile robot is taken as research platform in the paper. Robot stereo vision object tracking and obstacle avoidance based color image are studied in the paper. The content of the paper can be listed as follows:Firstly, development history and major content of robotic vision are introduced. And the content of robot stereo vision and current research status in domestic and abroad are explained emphatically.Secondly, the hardware and software platform of binocular vision system of Pioneer 3-DX mobile robot are introduced in detail. Key reaearch content of the paper is binocular stereo vision. So the principle,content and sorts of system of binocular stereo vision are expounded particularly.At last, both robot stereo vision object tracking and obstacle avoidance besed color image are expounded particularly. There are many defects if robot only depend color character to track object. The defects are improved in the paper. Robot proceed object tracking decision-making according to combine color character with shape character. Edge points and corners are taken as character match points in robot obstacle avoidance. And the distance between robot and obstacle according disparity image is calculated, then robot is steered to avoid obstacle. The experiment results show that the method possess better tracking effect by combining color character and shape character. The method improves defects that only depend color character. The obstacle avoidance method based Edge points and corners as character points possesses better robustness and anti-jamming. |
PDF Full Text Request