Hot-Rolled, Heavy-Rail Image Recognition Based on Deep-Learning Network

2021 ◽  
Author(s):  
Xie Changgui ◽  
Xu Hao ◽  
Liu Yuxi ◽  
Chen Ping
Keyword(s):  
Image Recognition ◽  
New Method ◽  
Processing Ability ◽  
Nonlinear Processing ◽  
Convolution Kernels ◽  
Defect Recognition ◽  
Deep Learning Network ◽  
Image Defect ◽  
Hot Rolled ◽  
Heavy Rail

A new method for image-defect recognition is proposed that is based on a convolution network with repeated stacking of small convolution kernels and a maximum pooling layer. By improving the speed and accuracy of image-defect recognition, this new method can be applied to image recognition such as heavy-rail images with high noise and many types of defects. The experimental results showed that the new algorithm effectively improved the accuracy of heavy-rail image-defect recognition. As evidenced by the simulation study, the proposed method has a lower error rate in heavy-rail image recognition than traditional algorithms, and the method may also be applied to defect recognition of nonlinear images under strong noise conditions. Its robustness and nonlinear processing ability are impressive, and the method is featured with high theoretical depth and important application value.

One Extended OTSU Flame Image Recognition Method Using RGBL and Stripe Segmentation

2011 ◽  
Vol 121-126 ◽  
pp. 2141-2145 ◽  
Author(s):  
Wei Gang Yan ◽  
Chang Jian Wang ◽  
Jin Guo
Keyword(s):  
Image Segmentation ◽  
Water Vapor ◽  
Image Recognition ◽  
Recognition Accuracy ◽  
New Method ◽  
Recognition Method ◽  
Otsu’S Method ◽  
Flame Shape ◽  
Image Segmentation Algorithm ◽  
Rgb Image

This paper proposes a new image segmentation algorithm to detect the flame image from video in enclosed compartment. In order to avoid the contamination of soot and water vapor, this method first employs the cubic root of four color channels to transform a RGB image to a pseudo-gray one. Then the latter is divided into many small stripes (child images) and OTSU is employed to perform child image segmentation. Lastly, these processed child images are reconstructed into a whole image. A computer program using OpenCV library is developed and the new method is compared with other commonly used methods such as edge detection and normal Otsu’s method. It is found that the new method has better performance in flame image recognition accuracy and can be used to obtain flame shape from experiment video with much noise.

scholarly journals A method for separating O-wave and X-wave and its application in digital ionosonde

2014 ◽  
Vol 56 (5) ◽  
Author(s):  
Wang Shun ◽  
Chen Ziwei ◽  
Zhang Feng ◽  
Gong Zhaoqian ◽  
Li Jutao ◽  
...  
Keyword(s):  
Image Recognition ◽  
Initial Phase ◽  
Phase Shifters ◽  
New Method ◽  
Digital Method ◽  
Electrical Method ◽  
Circularly Polarized ◽  
Automatic Scaling ◽  
Polarized Waves ◽  
Phase 0

<p><strong><em></em></strong>Separation for O wave and X wave is a very important job in interpretation of ionograms, which is premise for automatic scaling. In this paper, a new digital method for separating O wave and X wave is presented, based on a numerical synthesizing technique, which is different from using image recognition to separate trace O and trace X in the ionograms, and from using the electrical method to synthesize and detect circularly polarized waves. By replacing analog phase shifters and switches in existing ionosonde with digital phase shifters with different initial phase, 0°, +90°, −90°, circularly polarized waves are synthesized digitally within the range of 1-30 MHz, which eliminates the nonlinearity and expands the bandwidth of the ionosonde, and there is no need to switch the analog switches continuously. The new method has been successfully applied to CAS-DIS ionosonde and testing results show that the new digital method is capable of separating O wave and X wave well.</p>

Fast Position and Accurate Segmentation Algorithms for Detecting Surface Defects of the Thermal-State Heavy Rail Based on Machine Vision

2013 ◽  
Vol 5 (4) ◽  
pp. 40-60
Author(s):  
Xue Wang ◽  
Yiran Chen ◽  
Tao Cheng ◽  
Zhijiang Xie
Keyword(s):  
Nearest Neighbor ◽  
Surface Defects ◽  
Thermal State ◽  
Mean Shift ◽  
Sampling Point ◽  
Positioning Method ◽  
Rolled Condition ◽  
Novel Algorithms ◽  
Hot Rolled ◽  
Heavy Rail

Color imaging in the hot rolled condition provides the better reaction of heavy rail on surface defects. In this paper, it proposes a series of novel algorithms of accurate position and segmentation of surface defects of heavy rail. An image acquisition device is designed on the adjustable camera bracket with the linear array CCD, and based on the correlation among pixels at the image level, a fast positioning method is developed for searching the Region Of Interesting (ROI) of the surface defects. Using the Mean-Shift image filtering algorithm which features multi-parameter kernel function, amendments to the sampling point weights and regional search with the nearest neighbor sampling points, accurate segmentation of the identification character is easily achieved by K-means clustering. Experiments show that this algorithm for the identification of the heavy rail surface defects is proven to be more rapid in testing the inclusions, cracks and oxide skin defects with a good promotional value.

Combustion Synthesis of Ni-Ti-X Shape Memory Alloys

1991 ◽  
Vol 246 ◽  
Author(s):  
John J. Moore ◽  
Ru Chun Yi
Keyword(s):  
Shape Memory Alloys ◽  
Shape Memory ◽  
Combustion Synthesis ◽  
Shape Memory Effect ◽  
Memory Effect ◽  
Liquid Product ◽  
New Method ◽  
Transition Temperatures ◽  
Hot Rolled

AbstractCombustion synthesis is an energetically favorable new method of producing Ni-Ti series shape memory alloys (SMA′s). The ΔH°f,,298/Cp298 ratio plays a key role, especially if a liquid product is required. Solidified TiNi SMA's were hot rolled into plates exhibiting the shape memory effect. Ms transition temperatures can be controlled from -78 to 460°C by substituting Fe or Pd for Ni.

scholarly journals Design of Multi-Receptive Field Fusion-Based Network for Surface Defect Inspection on Hot-Rolled Steel Strip Using Lightweight Dataset

2021 ◽  
Vol 11 (20) ◽  
pp. 9473
Author(s):  
Wei-Peng Tang ◽  
Sze-Teng Liong ◽  
Chih-Cheng Chen ◽  
Ming-Han Tsai ◽  
Ping-Cheng Hsieh ◽  
...  
Keyword(s):  
Receptive Field ◽  
Steel Strip ◽  
Motion Blur ◽  
Gaussian Pyramid ◽  
Inadequate Information ◽  
Recognition Result ◽  
Defect Recognition ◽  
Hot Rolled ◽  
Hot Rolled Steel ◽  
Hierarchical Features

With the advancement of industrial intelligence, defect recognition has become an indispensable part of facilitating surface quality in the steel manufacturing process. To assure product quality, most previous studies were typically trained with many defect samples. Nonetheless, a large quantity of defect samples is difficult to obtain, owing to the rare occurrence of defects. In general, deep learning-based methods underperformed as they have inherent limitations due to inadequate information, thereby restraining the application of models. In this study, a two-level Gaussian pyramid is applied to decompose raw data into different resolution levels simultaneously filtering the noises to acquire compact and representative features. Subsequently, a multi-receptive field fusion-based network (MRFFN) is developed to learn the hierarchical features and synthesize the respective prediction scores to form the final recognition result. As a result, the proposed method is capable of exhibiting an outstanding performance of 99.75% when trained using a lightweight dataset. In addition, the experiments conducted using the disturbance defect dataset showed the robustness of the proposed MRFFN against common noises and motion blur.

Sign in / Sign up

Export Citation Format

Share Document