Weld Image Recognition Algorithm Based on Deep Learning

2019 ◽  
Vol 34 (08) ◽  
pp. 2052004
Author(s):  
Yan Li ◽  
Miao Hu ◽  
Taiyong Wang
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Weld Seam ◽  
Detection Method ◽  
Automatic Detection ◽  
Recognition Algorithm ◽  
Image Texture ◽  
Detection Accuracy ◽  
Weld Defects ◽  
Weld Inspection

As an important part of metal processing, welding is widely used in industrial manufacturing activities, and its application scenarios are very extensive. Due to technical limitations, the welding process always unavoidably leaves weld defects. Weld defects are extremely hazardous, and the work used must be guaranteed to be defect-free, regardless of the field. However, manual weld inspection has subjective factors such as inefficiency and easy missed detection, and although some automatic weld inspection methods have appeared, these traditional methods still do not meet actual demand in terms of detection time and detection accuracy. Therefore, there is a need for a higher quality weld image automatic detection method to replace the manual method and the traditional automatic detection method. In view of the above, this paper proposes a weld seam image recognition algorithm based on deep learning. The Adam adaptive moment estimation algorithm is chosen as the backpropagation optimization algorithm to accelerate the training of convolutional neural networks and design an independent adaptive learning rate. Through the simulation of the collected 4500 tube images, the adaptive threshold-based method is used for weld seam extraction. The algorithm proposed in this paper is compared with the weld seam recognition method based on image texture feature value distribution (ITFVD) and the SUSAN-based weld defect target detection method. The results show that the proposed method can identify weld defects in a short time on different sizes of weld images, and can further detect the type of weld defects. In addition, the method in this paper is better than the other two methods in the false detection rate, recall rate and overall recognition accuracy, which shows that the experimental results have achieved the expected results.

Weld Seam Recognition Based on Wavelet Analysis Algorithm

2011 ◽  
Vol 480-481 ◽  
pp. 1593-1597
Author(s):  
You Rui Huang ◽  
Na Na Zhao
Keyword(s):  
Genetic Algorithm ◽  
Wavelet Analysis ◽  
Image Recognition ◽  
Weld Seam ◽  
Wavelet Coefficient ◽  
Recognition Algorithm ◽  
Seam Tracking ◽  
Vision Sensor ◽  
Immune Genetic Algorithm ◽  
Analysis Algorithm

Vision sensor for weld seam location and seam tracking is developed. It is difficult for realizing picture precise identification based on machine welding at present, the image recognition algorithm based on wavelet analysis is designed in this paper. It uses immune genetic algorithm to optimize wavelet coefficient, and extracts the edge of picture preferably. In the process of detecting edges, compared with the others, the algorithm can inhibit the interference of noises effectively and remove artifacts of edges, which make the detected edge clearly and carefully. The result shows that the wavelet analysis algorithm has better effect for weld seam recognition.

Semi supervised inspection algorithm of automatic packaging curve based on deep learning

2021 ◽  
pp. 1-10
Author(s):  
Yong He
Keyword(s):  
Deep Learning ◽  
Supervised Learning ◽  
Optimization Algorithm ◽  
Posterior Probability ◽  
Detection Method ◽  
Detection System ◽  
Experimental Results ◽  
Detection Accuracy ◽  
Data Set ◽  
Packaging Process

The current automatic packaging process is complex, requires high professional knowledge, poor universality, and difficult to apply in multi-objective and complex background. In view of this problem, automatic packaging optimization algorithm has been widely paid attention to. However, the traditional automatic packaging detection accuracy is low, the practicability is poor. Therefore, a semi-supervised detection method of automatic packaging curve based on deep learning and semi-supervised learning is proposed. Deep learning is used to extract features and posterior probability to classify unlabeled data. KDD CUP99 data set was used to verify the accuracy of the algorithm. Experimental results show that this method can effectively improve the performance of automatic packaging curve semi-supervised detection system.

A deep Residual U-Net algorithm for automatic detection and quantification of ascites on the abdominopelvic computed tomography acquired in the emergency department (Preprint)

2021 ◽  
Author(s):  
Hoon Ko ◽  
Jimi Huh ◽  
Kyung Won Kim ◽  
Heewon Chung ◽  
Yousun Ko ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Deep Learning ◽  
Automatic Detection ◽  
High Specificity ◽  
Ct Images ◽  
Free Fluid ◽  
Detection Accuracy ◽  
Abdominopelvic Ct ◽  
Segmentation Accuracy ◽  
Detection And Quantification

BACKGROUND Detection and quantification of intraabdominal free fluid (i.e., ascites) on computed tomography (CT) are essential processes to find emergent or urgent conditions in patients. In an emergent department, automatic detection and quantification of ascites will be beneficial. OBJECTIVE We aimed to develop an artificial intelligence (AI) algorithm for the automatic detection and quantification of ascites simultaneously using a single deep learning model (DLM). METHODS 2D deep learning models (DLMs) based on a deep residual U-Net, U-Net, bi-directional U-Net, and recurrent residual U-net were developed to segment areas of ascites on an abdominopelvic CT. Based on segmentation results, the DLMs detected ascites by classifying CT images into ascites images and non-ascites images. The AI algorithms were trained using 6,337 CT images from 160 subjects (80 with ascites and 80 without ascites) and tested using 1,635 CT images from 40 subjects (20 with ascites and 20 without ascites). The performance of AI algorithms was evaluated for diagnostic accuracy of ascites detection and for segmentation accuracy of ascites areas. Of these DLMs, we proposed an AI algorithm with the best performance. RESULTS The segmentation accuracy was the highest in the deep residual U-Net with a mean intersection over union (mIoU) value of 0.87, followed by U-Net, bi-directional U-Net, and recurrent residual U-net (mIoU values 0.80, 0.77, and 0.67, respectively). The detection accuracy was the highest in the deep residual U-net (0.96), followed by U-Net, bi-directional U-net, and recurrent residual U-net (0.90, 0.88, and 0.82, respectively). The deep residual U-net also achieved high sensitivity (0.96) and high specificity (0.96). CONCLUSIONS We propose the deep residual U-net-based AI algorithm for automatic detection and quantification of ascites on abdominopelvic CT scans, which provides excellent performance.

scholarly journals A Deep Learning-Aided Detection Method for FTN-Based NOMA

2020 ◽  
Vol 2020 ◽  
pp. 1-11
Author(s):  
Jianxiong Pan ◽  
Neng Ye ◽  
Aihua Wang ◽  
Xiangming Li
Keyword(s):  
Deep Learning ◽  
Computational Complexity ◽  
Smart City ◽  
Mean Squared Error ◽  
Detection Method ◽  
High Reliability ◽  
Detection Methods ◽  
Low Latency ◽  
Detection Accuracy ◽  
Minimum Mean Squared Error

The rapid booming of future smart city applications and Internet of things (IoT) has raised higher demands on the next-generation radio access technologies with respect to connection density, spectral efficiency (SE), transmission accuracy, and detection latency. Recently, faster-than-Nyquist (FTN) and nonorthogonal multiple access (NOMA) have been regarded as promising technologies to achieve higher SE and massive connections, respectively. In this paper, we aim to exploit the joint benefits of FTN and NOMA by superimposing multiple FTN-based transmission signals on the same physical recourses. Considering the complicated intra- and interuser interferences introduced by the proposed transmission scheme, the conventional detection methods suffer from high computational complexity. To this end, we develop a novel sliding-window detection method by incorporating the state-of-the-art deep learning (DL) technology. The data-driven offline training is first applied to derive a near-optimal receiver for FTN-based NOMA, which is deployed online to achieve high detection accuracy as well as low latency. Monte Carlo simulation results validate that the proposed detector achieves higher detection accuracy than minimum mean squared error-frequency domain equalization (MMSE-FDE) and can even approach the performance of the maximum likelihood-based receiver with greatly reduced computational complexity, which is suitable for IoT applications in smart city with low latency and high reliability requirements.

Research on Burning Zone Detection Method Based on Flame Image Recognition for Ceramic Roller Kiln

2014 ◽  
Vol 602-605 ◽  
pp. 1761-1767
Author(s):  
Yong Hong Zhu ◽  
Peng Li
Keyword(s):  
Image Recognition ◽  
Working Conditions ◽  
Detection Method ◽  
Firing Process ◽  
Detection Accuracy ◽  
Process Data ◽  
Fusion Algorithm ◽  
Roller Kiln ◽  
Burning Zone ◽  
Ceramic Roller

In the firing process of ceramic products, the sintering conditions vary from firing phase to firing phase. In different firing phases, flame texture changes obviously, so it can be used as a important parameter of burning zone identification for ceramic roller kiln. In this paper, both flame image recognition of simulating artificial-look-fire and multi-point temperature detection technology are used to detect burning zone working conditions of ceramic roller kiln so as to greatly improve detection accuracy. The key data fusion algorithm of PTCR-based point detection temperature and flame image recognition–based detection method of burning zone working condition for ceramic roller kiln are proposed. The temperature measurement experiment system scheme of ceramic roller kiln burning zone is also given. The system can fuse the key process data with flame image characteristics so as to get the comprehensive database used to judge burning zone working conditions and temperatures. In the end, The testing experiment was carried out. The experimental results show that the method proposed above is feasible and effective.

Research on Bolt Crack Detection Method Based on Image Processing

2013 ◽  
Vol 433-435 ◽  
pp. 426-429
Author(s):  
Jin Qiu Liu ◽  
Bing Fa Zhang ◽  
Yu Zeng Wang ◽  
Guang Ya Li ◽  
Jing Ru Han
Keyword(s):  
Image Processing ◽  
Edge Detection ◽  
Image Recognition ◽  
Crack Detection ◽  
Detection Method ◽  
Image Acquisition ◽  
Detection Methods ◽  
Detection Accuracy ◽  
Contact Detection ◽  
Fracture Identification

A method of non-contact detection of bolt fracture have serial steps as follows: First of all the required data is obtained through image acquisition, then through the edge detection, image recognition and other image processing on the image to get the bolt fracture identification results, finally the non-contact measurement bolt fracture is realized. Experiments show that bolt crack detection method based on image processing, compared with the traditional detection methods improve the efficiency of detection and improve the detection accuracy. The method for bolt crack detection is feasible.

Deep Learning for the Detection and Recognition of Rail Defects in Ultrasound B-Scan Images

2021 ◽  
pp. 036119812110215
Author(s):  
Zhengxing Chen ◽  
Qihang Wang ◽  
Kanghua Yang ◽  
Tianle Yu ◽  
Jidong Yao ◽  
...  
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Defect Detection ◽  
High Performance ◽  
Image Data ◽  
Small Scale ◽  
Detection Accuracy ◽  
Data Set ◽  
Detection Rates ◽  
B Scan Images

Rail defect detection is crucial to rail operations safety. Addressing the problem of high false alarm rates and missed detection rates in rail defect detection, this paper proposes a deep learning method using B-scan image recognition of rail defects with an improved YOLO (you only look once) V3 algorithm. Specifically, the developed model can automatically position a box in B-scan images and recognize EFBWs (electric flash butt welds), normal bolt holes, BHBs (bolt hole breaks), and SSCs (shells, spalling, or corrugation). First, the network structure of the YOLO V3 model is modified to enlarge the receptive field of the model, thus improving the detection accuracy of the model for small-scale objects. Second, B-scan image data are analyzed and standardized. Third, the initial training parameters of the improved YOLO V3 model are adjusted. Finally, the experiments are performed on 453 B-scan images as the test data set. Results show that the B-scan image recognition model based on the improved YOLO V3 algorithm reached high performance in its precision. Additionally, the detection accuracy and efficiency are improved compared with the original model and the final mean average precision can reach 87.41%.

A deep learning-based image recognition algorithm for fecal shape of domestic rabbits

2018 ◽  
Vol 32 (s1) ◽  
pp. 67-78 ◽  
Author(s):  
Jingjing DEMOLOMBE ◽  
Tali YUAN ◽  
Xiao ZHANG ◽  
Longfei SHAO ◽  
Liheng GONG ◽  
...  
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Recognition Algorithm ◽  
Domestic Rabbits
Sign in / Sign up

Export Citation Format

Share Document