Convolutional neural network-based photovoltaic glass defect classification method and device

Abstract

The invention discloses a convolutional neural network-based photovoltaic glass defect classification method and a convolutional neural network-based photovoltaic glass defect classification device. The method comprises the following steps of carrying out multi-angle and variable-illumination image acquisition on a plurality of defect samples to obtain a plurality of images; preprocessing the images to fuse the multi-channel information and generate a multi-channel-fused defect sample image; adopting a convolution neural network model which meets a preset condition, extracting a network according to defect category design features and constructing a feature extraction convolutional neural network; obtaining the number of layers of all-connected neural networks and the number of neurons ofeach layer, and constructing a feature classification network; under the condition that the cross entropy loss function is minimized, completing the training of the convolutional neural network; according to an input sample image, outputting a prediction result of a defect category through the trained convolutional neural network. Based on the method, the situation that training sets are insufficient can be effectively solved while the generalization ability and the prediction precision of the model are guaranteed. Meanwhile, the high classification precision can be achieved for a small amountof glass defect samples.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a photovoltaic glass defect classification method and device based on a convolutional neural network. Background technique [0002] Photovoltaic glass defect classification is a relatively new field. Since 2005, the output of photovoltaic embossed glass has grown rapidly with the popularity of solar photovoltaic cells. The defect problem of photovoltaic glass has become more and more obvious with the increase in sales. On the one hand, after the defective photovoltaic glass is made into components, the glass Defects will cause irreparable components to be degraded or scrapped, and the losses caused are far greater than the cost of photovoltaic glass. On the other hand, some defects may cause consequences such as glass bursting in the subsequent tempering process of adding anti-reflection coatings, which poses serious safety hazards to the production process. As mentione...

AI Technical Summary

Problems solved by technology

In traditional methods, features are manually extracted, thresholds are set, cascaded tree classification usually relies heavily on the distribution of defective samples, and the labor cost is quite high

All kinds of defects are relatively diverse, and there are certain differences in the same type of defects under the same lighting conditions, and some common features of morphology that may exist are not easy to extract, so it is difficult for traditional methods to intuitively summarize strong high-order coarse-grained features of images effective classification

Images