Defect Detection Of Solar Panels Based On Infrared Thermal Imaging

Solar energy is currently the most widely used in all kinds of new energy forms for power generation. Solar panel is the kernel of the solar power devices, which has inevitable defects produced in the process of production and installation; it will seriously affect the power generation efficiency of solar panels. Therefore, the defect detection of solar panels is an essential link. Infrared detection technology is one of the hot spot in the field of industrial flaw detection, which has no need to filter visible light, high speed detection, high accuracy, non-contact detection, etc. Based on the conclusion of existing defect detection technologies of solar panels, this dissertation puts forward a method, which is a kind of online infrared imaging technology for defect detection of solar panels. Based on infrared thermal imaging technology, using the principle of infrared detection, reasonably combined with image processing technology, this article has developed and designed a set of system which is based on infrared thermal imaging for defect detection of solar panels. The details include:First of all, based on the research literature at home and abroad, this dissertation discusses the background and significance for defect detection of solar panels;Next, analyzing and summarizing the related theories of solar panels, defect characteristics of infrared, and the application of thermal infrared imaging in the infrared detection technology, which including the theory of infrared radiation, infrared detection principle, the theory of infrared imaging technology, all kinds of influence factor analysis of the infrared thermal imager, and so on. In addition, outlining the concept of digital image and the commonly used algorithm.Then, according to the requirements of real-time detection, by means of the device selection, designing of simulative light source system, compiling test software, completing a set of defect detection system of solar panels, providing overall design scheme and technical route of the system, finishing test of the elements.Afterwards, according to the detecting process and the features of defect panels, designing the system software and optimizing image processing algorithm. Using stage filter to optimize and improve the image resolution; on the basis of the traditional median filtering, improving median filtering algorithm, and makes up the disadvantage of fuzzy at the boundary point for the traditional median filter in removing noise; In the image edge detection algorithm, providing an improved Canny algorithm of double threshold automatic recognition, effectively detecting the four common defects of black pieces, broken gates, hidden cracks and fragments; Applying K-L transform to extract the defect area of image and calculate the difference of the average grey value, complexity, aspect ratio and the edge curve curvature,At the same time, extracting the feature information by straight line and cu rve fitting, and also calculating the length and area. This dissertation improves the tra ditional hough transform, detecting and measuring the length of the curve by fast sep aration arc line, at last, classifying defects by their size, for large defects, this essay h as applied the method of minimum circumscribed rectangle area computation, on the other hand, for small defects, it is based on the grayscale morphology to detect, and then mark the defects.This dissertation has designed a defect detection system for solar panels, which is based on infrared thermal imaging technology, experiments show that this system can meet the design requirements of detecting target, detecting situations include black piece damage area reaches 80%, the length of more than 8 cm hidden cracks and broken gates, area of more than 25 cm2 fragments, and etc. Detection speed can reach 3 targets per second, false drop rate is below 5%. With advantages of high efficiency, reliable operation and high accuracy, this system also has good social efficiency and economic benefits.