The invention relates to the technical field of loading and stacking, in particular to a mobile mechanical arm loading and stacking
system which comprises the following steps: acquiring three-dimensional
point cloud data of a vehicle through an optical detection unit; recognizing the size, posture and parking deviation of the vehicle; generating a stacking scheme based on a
random forest or a
deep learning model; performing real-time stack shape
simulation by combining a Poisson
reconstruction algorithm; and the stacking
robot is controlled to perform staggered stacking of transverse and vertical packages, and the compartment space
utilization rate is maximized. Cooperative control of all the units is achieved through the central control platform, and the overall operation fluency is improved. Through cooperation of the parking guiding unit and the high-flexibility
robot, horizontal and angle parking deviation is automatically compensated; a
visual system is adopted, and an intelligent
algorithm is combined, so that the recognition precision and robustness in a complex scene are improved; a self-adaptive stacking
algorithm is introduced, transverse and vertical bag staggering is supported, a bag slipping stack shape is prevented, and the space
utilization rate and stability are improved; the optical detection unit quickly completes vehicle modeling, and efficient planning and execution are realized in combination with a cloud edge collaborative
algorithm.