Surface Defect Recognition of Solar Panel Based on Percolation-Based Image Processing and Serre Standard Model

During the production process of solar panels, it is inevitable to have some defects, such as cracks on the surface of solar panels due to extrusion or damage due to quality issues. This article improves the Serre standard model, which can simulate the ventral visual pathway with object recognition ability, based on the latest research progress and results of simulating biological visual mechanism models in computer vision, to improve the recognition effect of surface defects on solar panels. At the same time, a pre-processing scheme combining Gaussian Laplace operator operator and adaptive Wiener filter to remove noise spots is studied, and the local Gabor Binary Pattern Histogram Sequence (LGBPHS) features are obtained through pre-processing. The Percolation-Based image processing method for detecting obvious cracks was used to determine the location of the algorithm and the calculation results based on the improved standard model method. It mainly refers to the MAX value output by the C2 layer and the classification and identification results of whether there are cracks, and the crack location function is completed. The experimental results show that the proposed method has an accuracy rate of 98.86% in training and 98.64% in testing, and both the false detection rate and the missed detection rate do not exceed 1%. Therefore, the method proposed in the study has a high accuracy and can effectively identify the surface defects of solar panels.

Automated summary

This study improves the Serre standard model and proposes a method to effectively identify the surface defects of solar panels. By combining pre-processing and binary image processing, the features of solar panels are obtained and classified using the improved model, achieving high accuracy in defect recognition.