A Binocular Camera Synchronization Calibration Method Using Closed-Loop Hybrid Optimization for the Froth Flotation Process

Binocular camera metric calibration has been widely applied in 3D reconstruction in industrial environments. However, due to the limitations of the complex environment in froth flotation sites, the conventional approach of fixing the camera’s relative position beforehand and then moving it to the flotation site is not feasible. Furthermore, because there are some hardware or software delays between the binocular cameras in froth flotation process, the left and right cameras are usually unsynchronized, which makes the calibration result inaccurate. Therefore, this paper proposes a synchronous optimization method of feature point to enhance the performance of binocular calibration. Initially, this study demonstrates an innovative two-phase hybrid optimization algorithm, called hill climbing-based improved differential evolution algorithm, to optimize all feature matching points synchronously. Then, we estimate an initial imaging delay of the left and right cameras based on empirical knowledge to accelerate the convergence speed. Furthermore, we develop a closed-loop optimization strategy to manage the nonlinear variations in delay specific to binocular cameras, thereby enhancing the model’s robustness. Finally, we conduct the experiments in the laboratory and industrial flotation process. The results show that the proposed method can effectively solve the unsynchronized problem of binocular cameras and increase the accuracy of camera calibration.