Research On Key Technologies Of Electric Vehicle Identification In Elevator Based On PCNN

In recent years,fire accidents cause by charging electric vehicles in high-rise civil buildings have occurred frequently,seriously threatening the lives and properties of residents.In order to avoid the illegal entry of electric vehicles into high-rise buildings,installing surveillance in elevators to identify electric vehicles inside elevators is one of the feasible methods to prevent their entry.A pulse coupled neural network(Pulse Coupled Neural Network,PCNN)based model for recognizing electric vehicles inside elevators is proposed for the problem of recognizing electric vehicles inside elevators.Based on this model,the following three aspects of research are carried out.(1)Study the basic principle of PCNN,and propose a PCNN-based model for recognition of electric vehicles in elevators(EV-PCNN).In this model,firstly,PCNN is applied to implement segmentation in the captured image;subsequently,the affine invariant moment algorithm is applied to extract the electric vehicle features in it;finally,a classifier is built by combining rough set and support vector machine to realize the recognition of electric vehicles in the elevator background.(2)To address the problem that there are many parameters in the EV-PCNN model and it is difficult to find the optimal combination of parameters,a parameter finding approach based on Seagull Optimization algorithm(Seagull Optimization algorithm,SOA)is proposed.The algorithm generates random populations in the solution space,and uses cross-entropy as the fitness function to guide the populations to perform an intelligent search for the space where the optimal solution is located.Simulation experiments show that the algorithm is able to find a better combination of PCNN initial weight parameters,which effectively improves the accuracy of electric vehicle recognition in elevators.(3)A seagull optimization algorithm incorporating Cubic chaos and refractive backward learning(CRSOA)is proposed to address the problems of slow convergence,easy prematureness and low accuracy of SOA algorithm in solving higher dimensional problems.Firstly,the initialized population is implemented based on Cubic chaos.Subsequently,the nonlinear parameter A is applied to improve the search ability of the particles for the solution space.Later in the iteration,refractive backward learning of the optimal particles is introduced to enhance the exploration of new solutions around the current optimal particle.Finally,the CRSOA algorithm is used for the parameter search optimization of PCNN to achieve the recognition of battery cars in elevators.The experimental results show that CRSOA has fast convergence,high solution accuracy and strong robustness,which can effectively improve the accuracy of electric vehicle recognition in elevators.