基于机器视觉的多巷道自动化立体仓库货位透明化分配方法

针对立体仓库内有效货位，设计透明化分配方案时主要通过人眼识别的方式掌握货位状态，使得最终分配方案的出库效率较低。因此，引入机器视觉技术，设计了一种新的仓库货位透明化分配方法。运用ABC分类法，对多巷道自动化立体仓库内所有货位进行分区。通过工业相机采集仓库内货架图像，通过机器视觉技术对采集图像进行识别，全面掌握仓库内处于空闲状态的货位位置。以最高出库效率为目标，建立货位透明化分配数学模型，再通过决策树搜索算法，生成最优货位分配方案。实验结果显示：应用所提方法进行货位分配，与基于粒子群算法、基于商品关联度的分配方法相比，使得出库效率提升了24.53%、23.92%。

Transparent Allocation Method of Multi Lane Automated Three-dimensional Warehouse Based on Machine Vision

For the effective location in the three-dimensional warehouse, when designing the transparent allocation scheme, the location status is mainly mastered through human eye recognition, which makes the delivery efficiency of the final allocation scheme low. Therefore, the machine vision technology is introduced to design a new warehouse location transparent allocation method. The ABC classification method is used to partition all cargo spaces in the multi lane automated three-dimensional warehouse. The shelf images in the warehouse are collected by industrial cameras, and the collected images are identified by machine vision technology, so as to comprehensively grasp the location of the empty storage space in the warehouse. Aiming at the highest delivery efficiency, the mathematical model of location transparent allocation is established, and the optimal location allocation scheme is generated through the decision tree search algorithm. The experimental results show that the proposed method improves the outbound efficiency by 24.53% and 23.92% compared with the allocation method based on particle swarm optimization and commodity relevance.