Photovoltaic module picture cutting method based on full convolutional neural network
A convolutional neural network and photovoltaic module technology, applied in the field of computer vision, can solve problems such as difficult direct detection, poor robustness of battery cells, and large image pixel magnitude, achieving high precision, good robustness, and reliability strong effect
Inactive Publication Date: 2021-04-20
聚时科技(江苏)有限公司
View PDF3 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0007] The purpose of the present invention is to provide a photovoltaic module image cutting method based on a fully convolutional neural network, which solves the problems of large image pixels, difficulty in directly detecting defects, and poor robustness when extracting battery cells
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0035] In order to verify the performance of the present invention, an experiment was done on the industrial data of photovoltaic panels in this embodiment. The present invention is described in detail below in conjunction with accompanying drawing and specific embodiment:
[0036] 1) Obtain 300 photovoltaic panel sample images and labels corresponding to the photovoltaic panel sample images. The training set contains 200 pictures of photovoltaic panels, and the test set contains 100 pictures;
[0037] 2) Normalize the photovoltaic panel image by calculating the mean value μ and standard deviation σ of the photovoltaic panel image;
[0038] 3) According to the existing problems, set the network structure as two paths of encoding and decoding, and then set the network structure parameters and training parameters including the number of layers, the number of channels and the size of the convolutional kernel of the fully convolutional neural network, and the training of the full...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a photovoltaic module picture cutting method based on a full convolutional neural network, and relates to the technical field of computer vision. The method comprises the following steps: firstly, constructing a full convolutional network, and then optimizing the network on a divided and preprocessed data set; secondly, inputting the preprocessed photovoltaic cell panel image into a trained full convolutional neural network to obtain a plurality of image masks corresponding to the cell units to form a first mask image; separating the adhered units in the first mask through a morphological corrosion algorithm to obtain a second mask image; detecting the area of the missing mask in the second mask image, and complementing the area by using the image mask which is completely the same as the battery piece unit in size and shape to obtain a third mask image; and superposing the third mask image and the photovoltaic cell panel image to obtain a segmented image of the cell unit. Compared with the prior art, the method has the advantages of high precision, high reliability, good robustness and the like.
Description
technical field [0001] The invention relates to the technical field of computer vision, in particular to a method for cutting pictures of photovoltaic modules based on a fully convolutional neural network. Background technique [0002] Photovoltaic modules, as the core part of solar power generation systems, have received more and more attention in recent years. Due to the complex production process and manufacturing process, various defects such as missing corners, hidden cracks, virtual welding and broken grids are prone to occur in photovoltaic modules, which will affect the power generation efficiency and service life of photovoltaic modules. Therefore, it is necessary to detect defects in photovoltaic modules in time during the production process. However, in the process of automatic defect detection, photovoltaic module images usually consist of tens of millions of pixels, and defects sometimes only occupy a very small part of the image, making it very difficult to di...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06T7/11G06N3/04G06N3/08
CPCY04S10/50
Inventor 张越超罗长志
Owner 聚时科技(江苏)有限公司