Research Of Aviation Baggage Transportability Calculation Method Based On 3D Point Cloud

As an important part of smart airport service,self luggage check-in is also an important guarantee for convenient travel of passengers.The self-service baggage check-in system needs to detect the adaptability of the baggage put in by the passengers according to the civil aviation regulations and the requirements of the carrier,and the point cloud can describe the physical appearance information of the baggage directly,so it is of great significance to study the calculation method of the adaptability based on the point cloud.The main work of this paper includes:Firstly,The collection method of baggage point cloud and the fast index structure of point cloud is studied.The point cloud collected by multiple cameras is accurately spliced to get a complete bag point cloud,and the experiment is used to verify the splicing effect.Secondly,The detection method of air baggage tray and strip is studied,using the normal vector feature of tray as the detection template,and using the face feature constraints of the point cloud to be matched to detect the tray,using RANSANC method to optimize the matching strategy.The method of color constraint and texture constraint is used to detect the baggage slip,and the experiment is used to verify the detection effect of the algorithm.Thirdly,The network structure of point cloud classification for air baggage is studied,using the lightweight MLP structure to extracts the global characteristics of baggage,the network using the edge convolution structure composed of X-Conv operator to improves the geometric detection ability.Verifies the network classification ability with the actual data.Finally,The discrimination calculation of the number of baggage pieces based on the point cloud segmentation is studied.The point cloud clustering method is used to make the segmented training data set.Based on the training results of PointNet network and SGPN structure,the multi-component point cloud is segmented and the number of bags is detected.The accuracy of the algorithm is verified by the simulation of the distribution of multiple consignment points and the production of data sets.