Subway foreign matter detection method, apparatus, and equipment, and subway shielded gate system
A foreign object detection and subway technology, applied in the field of image processing, can solve the problems of low efficiency and accuracy of foreign object detection and recognition in subways, and achieve the effect of low efficiency and accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0064] see figure 1 It is a schematic flow chart of a subway foreign object detection method provided in Embodiment 1 of the present invention. The method includes the following steps:
[0065] S11. Read the subway foreign object training sample, and acquire a two-dimensional color image in the training sample, perform preprocessing on the two-dimensional color image to extract its features, and obtain a processed two-dimensional color image, wherein the two-dimensional color image The three-dimensional color images include pictures containing subway foreign objects and pictures not containing subway foreign objects, and the capacity of training samples in the embodiment of the present invention is not limited.
[0066] S12. Obtain the convolutional neural network parameters obtained after training the processed two-dimensional color image through the classifier, and establish a trained convolutional neural network model;
[0067] Specifically, the softmax classifier is prefe...
Embodiment 2
[0081] Referring to embodiment one of the present invention and figure 1 The specific process of steps S11 to S16 described in , and see figure 2 Corresponding to the second embodiment of the present invention figure 1 The schematic flow chart of the specific preprocessing of the training samples in the step S11 shown, figure 1 Step S11 specifically includes:
[0082] S111. Read the subway foreign object training sample, and acquire the two-dimensional color image in the training sample;
[0083] S112. According to the structural requirements of the convolutional neural network, compress the two-dimensional color image into a two-dimensional color image with 64*64 pixels;
[0084] S113. Perform image segmentation on the two-dimensional color image with 64*64 pixels to obtain 8*8 image sub-regions;
[0085] Specifically, the image is divided into image sub-regions with a size of 8*8. The segmentation method is to divide the image into 8*8 image blocks from the upper left ...
Embodiment 3
[0134] Corresponding to the subway foreign object detection method disclosed in Embodiment 1 and Embodiment 2 of the present invention, Embodiment 3 of the present invention also provides a subway foreign object detection device, see Figure 5 It is a schematic structural diagram of a subway foreign object detection device provided in Embodiment 3 of the present invention, and the device specifically includes:
[0135] The first processing module 501 is used to read the subway foreign object training sample, and acquire the two-dimensional color image in the training sample, perform preprocessing on the two-dimensional color image to extract its features, and obtain the processed two-dimensional color image , wherein the two-dimensional color image includes a picture containing subway foreign objects and a picture not containing subway foreign objects;
[0136] Establishing module 502, for obtaining the convolutional neural network parameters obtained after training the proces...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 