Research On The Optimization Method Of Airport Baggage Automatic Loading And Stacking Based On Machine Learning

With the increasing number of airport passengers,the airport baggage throughput also increases,which puts forward higher requirements for the airport baggage handling system.In the process of airport baggage being transported from the baggage sorting port to the aircraft,the baggage must be stacked in accordance with the principle of first come first put,and the size of the baggage is different,so the problem of airport baggage packing is different from the traditional one It is not only a NP hard problem,but also an online decision-making problem.To solve this problem,this paper proposes the concept of spatial discretization,which combines the sparse theory with the deep support vector machine model to build a learning model based on the deep sparse least squares support vector machine,so that the baggage loading system has a more intelligent learning ability and meets the requirements of airport baggage transportation.First of all,according to the realization goal and work requirements of the airport baggage loading system,a general scheme and basic system structure for the airport baggage stacking method are designed.Through the spatial discretization of baggage and baggage car,the relationship between different positions is parameterized,and the airport baggage stacking strategy is formulated.At the same time,in order to avoid the problem that the baggage can not be accurately placed in the predetermined position in the actual loading process,which affects the next baggage stacking,the algorithm of baggage interference check is proposed to avoid the collision and other unsafe factors during loading.Secondly,the problem of baggage loading is also an online decision-making problem,which requires the algorithm to make a correct decision in a short time.Therefore,an algorithm model combining multi-layer structure and deep sparse least squares support vector machine is proposed.This method constructs multi-layer support vector machine model to learn the potential rules between position parameters.At the same time,in order to solve the problems of calculation complexity and calculation time increase caused by multi-layer structure,the least squares support vector machine technology is used instead of the traditional support vector machine The sparse processing can effectively solve the problem of algorithm speed.Then,the vision system of the loading robot is studied,and the problems of the different sizes of the baggage and the random placement angle in the loading process of the airport baggage are put forward.The vision recognition and positioning method for the airport baggage loading robot is proposed to obtain the image information through the depth camera,and then carry out the filtering,noise reduction and other processing to obtain more obvious features,and establish the feature improvement based on the contour Hu moment At the same time,the size of baggage is measured to complete the identification and positioning of baggage.Finally,through computer simulation and experiment to verify the airport baggage loading system,through computer programming to verify the feasibility of the baggage stacking strategy,using dobot four degrees of freedom machine to simulate the airport baggage loading environment for experiment,the experimental results show that the stacking strategy can effectively improve the space utilization of the baggage car,the airport baggage loading system can improve the efficiency of baggage loading.