Image segmentation method and device, electronic equipment and readable storage medium

Abstract

The invention discloses an image segmentation method and apparatus, an electronic device and a readable storage medium. The method comprises the steps of obtaining a to-be-detected image; inputting the to-be-detected image into a preset CNN segmentation network, at least obtaining a first feature map and a second feature map through the CNN segmentation network, wherein the first feature map and the second feature map are output from different convolution layers in the CNN segmentation network; and respectively inputting the first feature map and the second feature map into a preset attentionmechanism-based LSTM segmentation network, thereby obtaining an image segmentation result. According to the scheme, the to-be-detected image is acquired, the at least two groups of feature maps are determined by using the semantic segmentation network model based on the CNN, and then the fault area is segmented by using the LSTM segmentation network based on the attention mechanism according to the at least two groups of feature maps, so that the fault detection and segmentation efficiency and precision are improved.

Description

technical field [0001] The present application relates to the technical field of image segmentation and detection, in particular to an image segmentation method, device, electronic equipment and readable storage medium. Background technique [0002] The detection of faulty or defective parts in the industrial field has always been a common concern. Traditional fault detection mainly relies on manual methods, which has the disadvantages of low efficiency and high false detection rate. In recent years, with the advent of the Industry 4.0 era and the wide application of deep learning, fault detection using deep learning has gradually become the mainstream method. A major requirement in fault detection of industrial components is to pinpoint the fault area. like Figure 5 An example diagram of a failed component is shown, Image 6 Shown is the calibration map of the faulty area of ​​the faulty component. According to the figure, the specific location and shape of the faulty a...

AI Technical Summary

Problems solved by technology

Figure 7 ~ Figure 9 The left image is the segmentation result using the segmentation network, and the right is the labeled ground truth. Comparing the above left and right images, it can be found that the segmentation results of the existing image segmentation network have segmentation blur, fault location and Inaccurate boundary shapes and missing tiny targets

Images