Multi-threshold Defect Segmentation And Identification With Cuckoo Swarm Optimization Algorithm For Smartphone PCBA

Smartphone Printed Circuit Board Assembly(PCBA)defect identification such as spillage and over-welding is an important guarantee for automated production lines,and the precise segmentation of defect areas is the prerequisite for improving the accuracy of defect type and defect level identification.In view of the density of PCBA components and the diversity of defect types,it is a reasonable idea to adopt multi-threshold segmentation technology,but the optimization choice of multi-threshold lacks effective adaptive method.In order to improve the accuracy of PCBA defect area segmentation in smartphones to ensure the correctness of defect identification,this paper focuses on the optimization of multi-threshold selection,carry out the study multi-threshold segmentation algorithm based on cuckoo swarm optimization algorithm for defect areas and the defect identification study based on local area effect.The main contents include:(1)The advantages and defects of Cuckoo Search algorithm in image segmentation.Four common group intelligent optimization algorithms are compared.Selecting multiple sets of basic test functions as target functions,judging the performance of the four algorithms by the fitness value and applying the four algorithms to the classical image segmentation,the comparative experimental results show that the Cuckoo Search algorithm is superior in others.There are obvious advantages in choice,but at the same time,there are some major defects of the Cuckoo Search algorithm,such as limited search range,easy to fall into local optimization and lack learning of good solutions.(2)The study of cuckoo swarm optimization algorithm.The Lévy flight strategy based on the Cauchy distribution and dimensional update generates the candidate solution,generates the interference solution through the Normal random number,selects the advantage solution from the candidate solution and the interference solution by using the Metropolis criterion of the Simulated Annealing algorithm,and strengthens the learning of good solutions by using the roulette and two-way selection random walk strategy to generate a new solution to replace the eliminated solution,so as to propose an improved cuckoo swarm optimization algorithm.The comparative experimental results show that the improved algorithm can effectively avoid the defects of the traditional Cuckoo Search algorithm with limited search range,easy to fall into local optimality and lack learning of good solution,and can effectively obtain the global optimal solution and have good robustness.(3)The study of cuckoo swarm optimization multi-threshold segmentation algorithm for smartphone PCBA defect.For both OTSU multi-threshold segmentation algorithm and SNIC?Information Entropy multi-threshold segmentation algorithm,which lack adaptiveness in multi-threshold selection,the cuckoo swarm optimization multi-threshold OTSU segmentation algorithm and cuckoo group optimization multi-threshold SNIC?Information Entropy segmentation algorithm are proposed,and applied to smartphone PCBA defect segmentation.Experimental results show that both new segmentation algorithms have significantly improved the segmentation accuracy and robustness,and the overall performance of the cuckoo swarm optimization multi-threshold SNIC?Information Entropy segmentation algorithm is better.(4)The study of smartphone PCBA defect identification based on local area effects.For the two defect level identification tasks of PCBA spillage and over-welding of smartphones,the defect area detection is achieved by using the template matching algorithm based on gray information,and the precision segmentation of defect area is realized by using the multi-threshold SNIC?Information Entropy segmentation algorithm of cuckoo swarm optimization.On this basis,the local area effect method is adopted to calculate the pixel area of the target defect area,and a smartphone PCBA defect grade identification method is proposed based on the PCBA defect level standard and the calibration parameters of the imaging system.