The detection of solder joint defect and solar panel orientation based on ELM and robust least square fitting

This paper investigates the detection methods of solder joint defect and solar panel orientation based on extreme learning machine (ELM) and robust least square fitting (RLSF). The work first adopts image processing techniques to preprocess the images of solder joint and solar panel, then applies ELM to recognize those defected solder joints, and take the RLSF algorithm to acquire the edge of solar panel. Solder joint image features, such as area, gravity, anisotropy, and inertial moment are extracted and input to ELM for defect recognition. Experimental results show that the approaches can get a recognition rate of over 96%. For the solder joints defect detection, we can acquire the accurate edge of solar panel. After solar panel edge is determined, the orientation parameters - position shift, deflection angle and edge lengths of solar panel can be easily obtained.