Weld Defect Images Classification with VGG16-Based Neural Network

Liu, Bin; Zhang, Xiaoyun; Gao, Zhiyong; Chen, Li

doi:10.1007/978-981-10-8108-8_20

Cited by 48 publications

(27 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…(2) Convolutional transformation of images. The convolutional network in this paper uses VGG-16 [8] network. The network has 13 convolutional layers and 3 fully connected layers, totaling 16 layers of networks.…”

Section: Experiencementioning

confidence: 99%

Application of Neural Network in Image Detection of Illegal Billboards

Jiang¹,

Feng²,

Dong³

2020

Proceedings of the International Academic Conference on Frontiers in Social Sciences and Management Innovation (IAFSM 2019)

View full text Add to dashboard Cite

In recent years, urban management is gradually developing towards informationization. There are corresponding processing mechanisms for illegal billboards and illegal parking in urban management, but the method of detecting illegal billboards based on machine vision is still under study. At present, the automatic detection method of illegal billboards based on machine vision is generally carried out under ideal conditions, and the experiment is carried out under a simple background. This paper takes a street in Hebei Province as the research object, and studies the illegal billboard area under natural conditions, and proposes to use the machine vision-based method to automatically detect the illegal billboards. In the surveillance image, according to the street scene detected by the computer. Image information enables status evaluation of billboard placement. The detection algorithm of this paper uses the multi-angle suggestion area to accurately locate the illegal billboards in the image, and mark the detected billboards. Compared with the direct detection of the billboard in the image, the interference of the background factor on the target area is removed, and the false detection rate is effectively reduced. In billboard detection, this paper proposes an improved region suggestion algorithm to generate higher quality candidate regions in an image for adapting to billboards of various poses under natural conditions. Experimental data shows that the algorithm has good adaptability to the detection of billboards in surveillance images.

show abstract

Section: Experiencementioning

confidence: 99%

Application of Neural Network in Image Detection of Illegal Billboards

Jiang¹,

Feng²,

Dong³

2020

Proceedings of the International Academic Conference on Frontiers in Social Sciences and Management Innovation (IAFSM 2019)

View full text Add to dashboard Cite

show abstract

“…(2) Convolutional transformation of images. The convolutional network in this paper uses VGG-16 [9] network. The network has 13 convolutional layers and 3 fully connected layers, totaling 16 layers of networks.…”

Section: Experiencementioning

confidence: 99%

“…At present, there are not many researches on the use of the Faster-RCNN model for vehicle identification. In [9], Shoaib Azam et al used Faster-RCNN to perform vehicle pose detection from four different orientations (front, back, left and right); Tianyu Tang et al in [10] The vehicle data set obtained by aerial photography is based on the vehicle detection research based on Faster-RCNN. In real life, whether it is road monitoring, highway tolls or parking lots, the source of information for vehicle images is obtained by a single camera.…”

Section: Introductionmentioning

confidence: 99%

The Researches Detection Method of Illegal Parking Based on Convolutional Neural Network

Jiang¹,

Feng²,

Li³

2020

Proceedings of the International Academic Conference on Frontiers in Social Sciences and Management Innovation (IAFSM 2019)

View full text Add to dashboard Cite

Violation of parking is a legal term. Our country's law stipulates that there should be no-stop signs and markings, road sections with isolation facilities between motor vehicles and non-motor vehicle lanes and sidewalks, and crosswalks and construction sites. No parking; Railway crossings, sharp bends, narrow roads with widths less than 4 meters, bridges, steep slopes, tunnels, and sections within 50 meters of the above locations are not allowed to stop. In recent years, urban management is gradually developing towards information There are corresponding treatment mechanisms for illegal parking in urban management, but the method of detecting illegal parking based on machine vision is still under study. This paper takes a street in Hebei Province as the research object, and studies the illegal parking under the surveillance image. It also proposes to use the machine vision-based method to automatically detect the illegal parking. In the monitoring image, according to the street view image information detected by the computer, A status assessment of the vehicle placement can be achieved. The detection algorithm in this paper uses the multi-angle suggestion area to accurately locate the offending vehicles in the image and mark them on the detected vehicles. The experimental data shows that the algorithm has good adaptability to the detection of illegal parking in surveillance images, and effectively improves the inspection efficiency.

show abstract

“…The purpose is to ensure that the main information of the image is extracted while minimizing the occurrence of noise and interference and improving the classification accuracy. In Reference [19], Liu Bin et al utilized the VGG16 based fully CNN to inspect welding defect images using the idea of transfer learning. High precision classification is achieved with a relatively small data set.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning-Based Classification of Weld Surface Defects

Zhu

Liu

2019

Applied Sciences

View full text Add to dashboard Cite

In order to realize the non-destructive intelligent identification of weld surface defects, an intelligent recognition method based on deep learning is proposed, which is mainly formed by convolutional neural network (CNN) and forest random. First, the high-level features are automatically learned through the CNN. Random forest is trained with extracted high-level features to predict the classification results. Secondly, the weld surface defects images are collected and preprocessed by image enhancement and threshold segmentation. A database of weld surface defects is established using pre-processed images. Finally, comparative experiments are performed on the weld surface defects database. The results show that the accuracy of the method combined with CNN and random forest can reach 0.9875, and it also demonstrates the method is effective and practical.

show abstract

Weld Defect Images Classification with VGG16-Based Neural Network

Cited by 48 publications

References 10 publications

Application of Neural Network in Image Detection of Illegal Billboards

Application of Neural Network in Image Detection of Illegal Billboards

The Researches Detection Method of Illegal Parking Based on Convolutional Neural Network

Deep Learning-Based Classification of Weld Surface Defects

Contact Info

Product

Resources

About