Research On Quality Detection Algorithm Of Solar Cells Based On Image

The traditional solar cell manufacturing process relies on manual detection of defects,but manual detection is inefficient and costly.With the rapid development of artificial intelligence,intelligence is the future trend of industrial inspection.The image processing technology is used to analyze the solar cell image and detect the quality of the solar cell quickly and accurately.It has strong theoretical research significance and practical application value.This paper firstly summarizes the types of solar cell defects that can be detected under natural light conditions.Secondly,for the problem that the obtained images are distorted when solar cells are randomly placed in actual production,this paper proposes an image correction algorithm based on a multi-step search algorithm that corrects the image and extracts the corresponding region of interest,which solves the problem of low robustness,low accuracy,and low real-time performance of the conventional correction algorithm.For the problem that the detection of defects is difficult due to the complex texture of the solar cell surface,this paper proposes a corresponding surface texture removal algorithm to solve this problem.Aiming at the problem that the traditional defect detection algorithm detects a small number of defects,especially when multiple defects are present at the same time,this paper proposes a set of tests for specific defects(notches,collapses,holes,dirt,scratches).The algorithm solves the problem that the traditional algorithm detects single defects.For the geometric measurement in traditional algorithms,only the deflection angle is taken into consideration.In this paper,the offset and gate position are added to meet the needs in the actual production process,which solves the problem of less traditional geometric measurement range.Finally,based on the above algorithm and combined with hardware and software,a complete solar cell quality detection system was designed and implemented.Finally,the feasibility of the system was proved through experiments.Experimental results show that compared with the traditional quality detection algorithms,the correction success rate and calculation speed of the algorithm proposed in this paper have been improved;the detection accuracy of solar cell defects has been improved and the robustness has been properly enhanced,especially in When multiple defects exist at the same time;for the geometric measurement of solar cells,the real-time performance is improved under the same accuracy.