Bearing surface deformation and character mixed defect image detection method

Abstract

The invention discloses a bearing surface deformation and character mixed defect image detection method. The method mainly comprises the following steps: constructing a text detection model and a defect segmentation model, constructing a bearing image data set, a text label data set and a defect label data set by using existing image data, and training the text detection model and the defect segmentation model to obtain two trained text detection models and defect segmentation models; building a hardware platform, and shooting the surface of the to-be-measured bearing by using the built hardware platform and the industrial camera to obtain a bearing surface picture; and inputting a to-be-detected bearing surface picture, performing fusion processing according to the text detection model and the defect segmentation model, and outputting a detection result. According to the invention, multi-neural-network-fused image detection is used, so that the detection precision of the deformation defect of the pit on the surface of the bearing is improved, the problem that defects and characters cannot be distinguished by single traditional vision is solved, and the problem that the defects aredifficult to detect in a character area by a single neural network is solved.

Description

technical field [0001] The invention relates to a bearing surface image processing method, in particular to a bearing surface deformation and text mixed defect image detection method. Background technique [0002] Bearings are the most widely used and most important parts in the machinery industry. Whether it is precision, mass production, surface quality, or machining accuracy, they all have a huge impact on bearing life. However, in the actual production process, due to the influence of the enterprise's technology, the bearing surface will cause various damages, and the most important deformation defect is the deformation of the bearing surface pit. Therefore, it is possible to detect whether there are pits and deformations on the bearing surface, which can bring positive significance to the production of bearings for enterprises. [0003] Although there are currently a wide variety of bearing defect detection methods, most of these methods are used in the diagnosis of be...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problems existing in the background technology, the present invention mainly provides an image detection method for bearing surface deformation and text mixed defects, which solves the problems existing in the prior art when the existing neural network is used to identify the defects on the bearing surface and the outer side of the bearing. The problem of the influence of the text on the recognition result, and the problem of poor performance and low confidence when detecting the defect part overlapping with the text area

[0009] The text detection model uses a fully convolutional network with convolution kernels of different lengths and widths to solve the problem of recognizing and distinguishing the shape of text strips; the defect segmentation model uses the pixel-level classification characteristics of the segmentation network to classify the entire image

Images