Design of Image Based Optical Flow Tracking and Positioning System in Intelligent Assembly

The movement of workers in intelligent assisted assembly scenarios can cause model shaking, as well as issues such as obstruction, severe tilting, and scale changes of landmarks, which can lead to failed tracking of landmarks. Based on this, a high-precision and fast detection and tracking method is proposed. Firstly, the framework of the marker detection and tracking algorithm was established, and its working principle and process were explained; Secondly, coordinates are provided for subsequent algorithms through marker detection calculations; Then, the pyramid optical flow algorithm in the visual odometer is added to process the coordinates, and a tracking effect is added to the landmark coordinates to combat the problems caused by worker movement; Finally, when using the processed coordinates for pose solving, a nonlinear optimization matrix that combines the two-step degree method and Gaussian Newton method is added to enhance the algorithm’s anti-interference ability. The simulation experiment results show that the marker detection and tracking algorithm runs quickly, with guaranteed tracking accuracy and efficiency. It can effectively solve related problems such as image jitter, occlusion, and scale changes, and effectively reduce the impact of lighting. The design has achieved the expected results.