An automatic layout method for railway station signal equipment drawings

Abstract

This invention discloses an automatic layout method for railway station signaling equipment drawings. Specifically, it involves: converting a railway station track layout drawing to be designed into an image format; inputting the obtained railway station track image into a Color-block GAN; the Color-block GAN automatically selecting the locations where various types of railway signaling equipment should appear based on the track information and railway technical specifications in the input image, obtaining an output image containing track and equipment location information; reading the equipment category and location from the output image; and automatically completing the equipment layout on the layout drawing to be designed based on the read equipment category and location information. This invention solves the problem of existing methods requiring a large amount of repetitive manual operation, and can automatically complete the layout of railway station signaling equipment, reducing the workload of designers in the preliminary design stage.

Description

Technical Field

[0001] This invention belongs to the field of railway drawing design, and specifically relates to an automatic layout method for railway station signal equipment drawings. Background Technology

[0002] The design of railway station signaling equipment drawings is one of the key issues in railway construction, aiming to arrange the signaling equipment in the track layout drawings. In engineering practice, the design of railway signaling equipment layout drawings can be divided into two parts: preliminary design and construction drawing design. The preliminary design stage needs to determine the engineering design scheme, while the construction drawing design stage needs to provide detailed construction drawings, including specifications and all drawings. To accelerate this design process, computer-aided design has been introduced into this field.

[0003] Existing computer-aided layout methods for railway drawings primarily simplify the process through software extensions and secondary development. For example, plugins are developed for commonly used drafting software like CAD, modularizing frequently used steps in railway drawing layout to make the process simpler and faster for designers. While this method effectively improves designer efficiency, the sheer number of railway drawings required and many repetitive and simple steps still necessitate manual operation, resulting in significant repetitive work.

[0004] To further improve design speed, researchers began to study automated layout methods. Some researchers focused on the construction drawing design phase of railway signaling equipment layout, manually collecting basic station data to create various information tables, thereby accurately calculating and drawing the plans. However, because the construction drawing design phase requires detailed drawings, this method involves manual tasks such as setting points and reference lines on the drawings, making it quite complex.

[0005] This invention primarily studies the preliminary design stage of railway signaling equipment layout design, which can be divided into three steps: initial draft design, on-site verification, and design refinement. The initial draft design mainly provides a standard signaling equipment layout diagram based on railway technical specifications, serving as the foundation for subsequent on-site verification and adjustments. Completing this part of the work quickly will allow designers to better focus their efforts on the subsequent on-site verification and analysis. Therefore, it is necessary to expedite the completion of this universally applicable initial draft design stage. Summary of the Invention

[0006] To address the aforementioned problems, this invention provides an automatic layout method for railway station signal equipment drawings.

[0007] The present invention provides an automatic layout method for railway station signal equipment drawings, comprising the following steps:

[0008] Step 1: Convert a railway station track layout diagram to be designed into an image format.

[0009] Step 2: Input the obtained railway station track image into Color-block GAN.

[0010] Step 3: Color-block GAN automatically selects the locations where various railway signaling devices should appear based on the track information and railway technical specifications in the input image, and obtains an output image containing track and device location information.

[0011] Step 4: Read the device category and location from the output image.

[0012] Step 5: Automatically complete the equipment layout on the layout diagram to be designed based on the obtained equipment category and location information.

[0013] Furthermore, the Color-block GAN network is an improved conditional generative adversarial network, comprising a generator and a discriminator network. The generator network consists of convolutional downsampling, six residual blocks, and deconvolutional upsampling; the discriminator network consists of several convolutional downsampling layers. Color-block GAN also includes a category decoupling module, which separates different categories of railway signaling equipment in a hierarchical manner, thereby learning their distribution characteristics according to category.

[0014] Furthermore, during training, Color-block GAN fixes one network parameter in the generator and the discriminator while updating the other network parameter, repeating this process in an adversarial manner until the generation effect meets the requirements. When updating the generator network while fixing the discriminator network, the discriminator network, L1 loss, and class loss function are considered simultaneously. The class loss function incorporates some railway technical specifications, mainly concerning direction, scale, and the relative positional relationship between equipment.

[0015] Furthermore, during automatic placement, only the trained generator network is used.

[0016] The beneficial technical effects of this invention are as follows:

[0017] This invention can solve the problem of existing methods requiring a large amount of repetitive manual operation, and can automatically complete the layout of railway station signal equipment, reducing the workload of designers in the preliminary design stage. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the automatic placement method of the present invention.

[0019] Figure 2This is a schematic diagram of the processing procedure after an image is input into a Color-block GAN in an embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram illustrating the process of updating the Color-block GAN generator network in an embodiment of the present invention. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0022] The flowchart of an automatic layout method for railway station signaling equipment drawings according to the present invention is as follows: Figure 1 As shown, specifically:

[0023] Step 1: Convert a railway station track layout diagram to be designed into an image format.

[0024] Step 2: Input the obtained railway station track image into Color-block GAN.

[0025] Step 3: Color-block GAN automatically selects the locations where various railway signaling devices should appear based on the input image, and obtains an output image containing track and device location information.

[0026] Step 4: Read the device category and location from the output image.

[0027] Step 5: Automatically complete the equipment layout on the layout diagram to be designed based on the obtained equipment category and location information.

[0028] In step 1, the railway track image is obtained by cropping the area containing the railway station from the electronic drawing.

[0029] In step 2, the railway station track image is input into the Color-block GAN, and the processing procedure is described below. Figure 1 The main steps are to layer the input images according to device category, input them into the generator network, process them and output them according to category, and then merge and output them.

[0030] In step 2, the Color-block GAN network is an improved conditional generative adversarial network, thus containing two networks: a generator and a discriminator. The main improvement of this invention lies in: 1. Simplifying railway signaling equipment, retaining only location and category information. Figure 1 1. The position of the colored block represents the location of the equipment, and the difference in the color block filling represents the equipment category; 2. The network learns the location patterns of various types of equipment separately; 3. Stronger positional relationship constraints are added to the design of the loss function, making the generated equipment locations more consistent with railway technical specifications regarding direction, scale, and relative positional relationships between equipment.

[0031] During training, a large number of paired images of railway track and signaling equipment are input into the network structure as a dataset, enabling it to learn the mapping relationship between the railway track images and the signaling equipment images, until the network can automatically select the location of the railway signaling equipment effectively. The process of updating the generator network is described in [link to documentation]. Figure 3 Where x is the n-layer track line image, output G(x) is the n-layer generated image, and y is the n-layer target image. The L1 loss function is calculated using formula (1). The category loss function is calculated using formula (2). This loss function mainly considers the directional and relative positional relationships between different equipment categories in railway technical specifications. During training, it reduces the differences between the generated image and the real image in these aspects, making the generated image more in line with the specifications. This represents the coordinates of the color patch at the center of the i-th layer of the target image y. Similarly.

[0032]

[0033]

[0034] In step 3, the output of the generator network is obtained, which is the output image containing track and device location information.

[0035] In step 4, by Figure 2 As can be seen, the output image is layered according to the type of signal device, and the device location can be obtained by detecting each layer (i.e. each type) of the image.

[0036] In step 5, the obtained equipment category and location information are converted into category and coordinate information in the drawing, and the layout is completed on the layout drawing by calling pre-edited blocks.

[0037] This method can solve the problem of existing methods requiring a lot of repetitive manual operations, and can automatically complete the layout of railway station signal equipment, reducing the workload of designers in the preliminary design stage.

[0038] The scope of application of this invention is not limited to the design of railway station signaling equipment drawings.

Claims

1. An automatic layout method for railway station signal equipment drawings, characterized in that, Includes the following steps: Step 1: Convert a railway station track layout drawing to be designed into an image format; Step 2: Input the obtained railway station track image into Color-block GAN; The Color-block GAN network is an improved conditional generative adversarial network, which includes two networks: a generator network and a discriminator network. The generator network consists of several parts, including convolutional downsampling, six residual blocks, and deconvolutional upsampling. The discriminator network consists of several convolutional downsampling layers. The Color-block GAN also includes a category decoupling module, which separates different categories of railway signaling equipment in a hierarchical manner, thereby learning their distribution characteristics according to categories. During training, the Color-block GAN uses one network parameter fixed and the other network parameter updated in a loop to train against each other until the generation effect meets the requirements. When updating the generator network while fixing the discriminator network, the discriminator network, L1 loss and class loss function are considered simultaneously, with the class loss function incorporating some railway technical specifications. During training, a large number of paired images of railway track and railway signal equipment are used as a dataset and input into the network structure to learn the mapping relationship between railway track images and railway signal equipment images until the network can automatically select the location of railway signal equipment effectively. Update the generator network, where x is the n-layer orbital image, the output G(x) is the n-layer generated image, y is the n-layer target image, and the L1 loss function is calculated using formula (1): （1） The formula for calculating the category loss function is formula (2): （2） This loss function primarily considers the directional and relative positional relationships between different equipment categories in railway technical specifications. During training, it reduces the differences between generated and real images in these aspects, making the generated images more compliant with the specifications. , ) represents the coordinates of the color patch at the center of the i-th layer of the target image y. , Similarly; Step 3: Color-block GAN automatically selects the locations where various railway signaling devices should appear based on the track information and railway technical specifications in the input image, and obtains an output image containing track and device location information; Step 4: Read the device category and location from the output image; Step 5: Automatically complete the equipment layout on the layout diagram to be designed based on the obtained equipment category and location information.

2. The automatic layout method for railway station signal equipment drawings according to claim 1, characterized in that, During automatic placement, only the trained generator network is used.

Images