Detection methods for micro-cracked defects of photovoltaic modules based on machine vision

The efficiency and the service life of the photovoltaic modules are affected by the surface defects. Therefore, it is critical to detect the photovoltaic modules whether it is qualified or not before assembling into solar panels. This paper applies a method to detect micro-cracked defects in photovoltaic modules using electroluminescence (EL) technology and image processing. After applying forward bias voltage to photovoltaic modules, a large amount of non-equilibrium carriers is injected into photovoltaic modules from the diffusion region recombination to constantly composite luminescence and emit photons. Then an image is formed by a CCD camera which is used to capture these photons. As the brightness of the captured image is proportional to minority carrier diffusion length and current density, if the minority carrier diffusion length is relatively low, there may be defective, which results in a relatively dark image. Micro-cracked defects can be effectively found by analyzing the EL image. Varies of methods, including image segmentation, Gauss filtering, Hough line detection, are used to process image to judge whether the solar cell module is cracked. According to detecting results, the combination of these methods can effectively detect micro-cracked defects in photovoltaic modules.