Laser curve extraction of a train wheelset based on an encoder-decoder network

Abstract

An algorithm of laser curve segmentation for a train wheelset based on an encoder- decoder network is proposed. Aiming at the rich local features and simple semantic features of the train wheelset laser curve image, a neural network with shallow depth, high resolution, and good detail performance was designed. The proposed neural network makes full use of the dense connection mechanism and the upsampling module to enhance feature reuse and feature propagation. It can extract context semantic information at multiple scales with fewer parameters. Experimental results show that the encoder-decoder network has better performance than other neural networks in laser curve extraction of train wheelset. Based on the encoder-decoder neural network, mIOU, Recall, Accuracy, and F1_score of the laser curve dataset of the train wheelset, the score index reached 86.5%, 89.2%, 99.9%, and 85.0%, which can accurately extract the laser stripe of the train wheelset. Additionally, the encoder-decoder network can diminish the influence of noise on the extraction of laser fringes of a train wheelset to a certain extent. Therefore, it has good application in railway safety.

Laser Curve Extraction of Wheelset Based on Deep Learning Skeleton Extraction Network

“…Although many scholars have tried to introduce deep learning methods, most of them are used for denoising and cannot directly extract the center of the laser stripe. For example, Wang Shengchun et al [ 8 ] used deep learning E-Net for preprocessing, followed by template matching, and the gray gravity center method for subsequent processing, and Yang Kai et al [ 9 ] used UNet network for laser fringe segmentation. The results of these methods often depend on the effect of deep learning preprocessing and consume a great deal of time.…”

Research on adaptive image enhancement algorithm in laser stripe extraction

“…In recent years, image processing technology has been widely used in the field of workpiece dimensional measurement due to the increasing requirements for metal workpiece dimensional measurement accuracy in engineering applications and the rapid development of image processing technology [1] . In the line structured light measurement system, the system light source projects laser stripes onto the workpiece surface to form laser stripes characterizing the workpiece dimensional information, and the camera acquires the laser stripe images, and then obtains the workpiece dimensional information by image analysis of the acquired laser stripe images [2] . In the process of image acquisition, due to the system light source, ambient light and reflection of the workpiece, the collected laser stripe images may have the problem of uneven illumination, which is locally too bright or too dark, which affects the accuracy of laser stripe features extraction [3] .…”