Photovoltaic cell color difference classification method based on color segmentation

Abstract

The invention provides a photovoltaic cell color difference classification method based on color segmentation. The method comprises the steps of acquiring an original image of the photovoltaic cell panel, inputting the original image into a key point detection network to detect key points of the photovoltaic cell panel, and outputting a key point thermodynamic diagram and key point coordinates; connecting the key points in sequence to obtain a photovoltaic cell panel area image, and carrying out projection transformation on the photovoltaic cell panel area image to obtain a photovoltaic cell panel orthophoto image ImageA; detecting impurities in ImageA through an impurity detection network, converting the impurities into white and outputting ImageB; setting a plurality of reference colors,classifying pixels based on the reference colors and KDTree nearest neighbor search, outputting a color classification image ImageC, and segmenting the color classification image ImageC into a plurality of photovoltaic cell panel grid region images according to a certain proportion; and counting the proportion K of all reference color pixels in the total pixel number in each grid region image, and judging whether the photovoltaic cell panel grid region has color difference.

Description

technical field [0001] This application relates to the field of image processing, in particular to a method for color difference classification of photovoltaic cells based on color segmentation. Background technique [0002] The color of photovoltaic panels is mainly affected by the texturing process in the production process. The poor types of photovoltaic panels include flower slices, red slices and some yin and yang slices. At this stage, it mainly relies on manual observation to judge whether there is color difference in photovoltaic panels, which has the problems of long time consumption, low efficiency, high cost, and inconsistent judgment standards. Moreover, long-term human eye sorting will inevitably cause human eye fatigue, which will lead to a decrease in work efficiency and an increase in false detection rates. Contents of the invention [0003] In view of the above problems, the present invention proposes a method for color difference classification of photov...

AI Technical Summary

Problems solved by technology

At this stage, it mainly relies on manual observation to judge whether there is color difference in photovoltaic panels, which has the problems of long time consumption, low efficiency, high cost, and inconsistent judgment standards

Moreover, long-term human eye sorting will inevitably cause human eye fatigue, which will lead to a decrease in work efficiency and an increase in false detection rates.

Images