Camera overexposure anomaly adjustment method and system for photovoltaic cell panel image acquisition

Abstract

The invention provides a camera overexposure anomaly adjustment method and system for photovoltaic cell panel image acquisition, and relates to the field of artificial intelligence. The method comprises the steps of collecting a photovoltaic cell panel sample image and extracting pixel point brightness, if the pixel point brightness is greater than a first threshold, determining that overexposureis suspected; preprocessing the sample image to obtain a photovoltaic cell panel image, carrying out contour recognition on the photovoltaic cell panel image, determining whether the image is speckle-like overexposure, and for the photovoltaic cell panel image which does not accord with speckle-like overexposure conditions, obtaining a binarized region segmentation map; and if the total number ofregion pixel points greater than a brightness threshold is not less than a preset area threshold, determining that the region pixel points are light erosion type overexposure; and analyzing the overexposure type, the area coefficient, the contrast ratio and the influence degree of the overexposure image, establishing an overexposure anomaly adjustment model to adjust camera parameters and poses, and matching the optimal exposure degree. According to the invention, the overexposure type can be predicted, the exposure can be adjusted in a targeted manner, camera parameters and poses can be adjusted in a self-adaptive manner according to different environments, and the overexposure adjustment precision can be improved.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to a camera overexposure abnormality adjustment method and system for collecting images of photovoltaic panels. Background technique [0002] The photovoltaic industry has received great attention due to its clean and sustainable characteristics. The most important part of the photovoltaic industry is intelligent operation and maintenance, which implements automation and intelligent operation and maintenance while reducing costs as much as possible, so as to obtain maximum benefits. At this stage, in the defect detection of photovoltaic panels, various cameras, cameras and other image acquisition devices are used to support the data perception part. The cameras are often located in a situation with sufficient sunlight, so it is easy to cause the camera to be overexposed, which is very inconvenient for subsequent image processing. [0003] At this stage, there are some simple...

AI Technical Summary

Problems solved by technology

[0004] This method is universal, but for the specific type of photovoltaic cell camera, the environment where the photovoltaic cell panel is located is sufficiently illuminated, and the photovoltaic cell panel itself has welding strips, edges, and grid lines. In the process of exposure adjustment, the accuracy of this method is not high, and it is easy to cause misjudgment and misjudgment.

Images