Recognition of industrial machine parts based on transfer learning with convolutional neural network

Li, Qiaoyang; Chen, Guiming

doi:10.1371/journal.pone.0245735

Cited by 4 publications

(5 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Cheng et al combined integration learning with random cropping and developed a random cropping integrated neural network (RCE-NN) to overcome the complex background environment problem [27]. Li et al proposed an algorithm capable of accurate classification of parts, that is, a convolutional neural network model based on the InceptionNet-V3 pre-trained model through migration learning and reached an accuracy of 99.74% in the test set, and the algorithm can be applied in the intelligent diagnosis and maintenance of the parts with good prospects [28].…”

Section: Industrial Part Classificationmentioning

confidence: 99%

“…Li et al. proposed an algorithm capable of accurate classification of parts, that is, a convolutional neural network model based on the InceptionNet‐V3 pre‐trained model through migration learning and reached an accuracy of 99.74% in the test set, and the algorithm can be applied in the intelligent diagnosis and maintenance of the parts with good prospects [28].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A DenseNet‐based feature weighting convolutional network recognition model and its application in industrial part classification

An,

Sun,

Song

et al. 2023

IET Image Processing

View full text Add to dashboard Cite

Traditional warehousing typically needs machine learning or manual tagging to classify objects. However, this method is less robust and consumes a lot of labour and material resources. Based on DenseNet, this work proposes a feature weighting convolutional network recognition model and designs a set of software and hardware for data acquisition, which is applied to the efficient classification of industrial parts in warehouse management. Firstly, this work modifies DenseNet by embedding SE‐Block, and replaces the cross‐entropy loss function with the focus loss function to optimize the model structure. Secondly, a multi‐view hardware and software acquisition system is designed to complete the functions of part image acquisition, image preprocessing, model training and part recognition. Finally, an industrial parts sorting experiment was designed. Compared with the original DenseNet model, the proposed weighted convolutional network identification model showed that the accuracy of the modified model was increased by 3.09% and the convergence rate was significantly improved. The modified model proposed in this work aims to improve the recognition accuracy of industrial parts in modern warehouse management, so as to modify the classification efficiency of warehouse parts in production.

show abstract

Section: Industrial Part Classificationmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

A DenseNet‐based feature weighting convolutional network recognition model and its application in industrial part classification

An,

Sun,

Song

et al. 2023

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…The finer the particles of LSCT, the better the fluidity of the concrete during pouring, the larger the surface area of the LSCT particles participating in the reaction, and the stronger the activity [25]. However, it is easy to increase the consistency of the concrete slurry and cause insufficient gas generation, and the larger the particle size of LSCT, the worse the fluidity of the concrete and subsidence [26]. Therefore, it is very important to choose LSCT of appropriate fineness to prepare concrete.…”

Section: The Effects Of Mechanical Force On the Particle Size Andmentioning

confidence: 99%

Image Analysis Technology in the Detection of Particle Size Distribution and the Activity Effect of Low‐Silicon Copper Tailings

Zhao

Liu

et al. 2021

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

To speed up the comprehensive utilization and treatment of copper tailings, the digital image processing technology is proposed in this study to detect the low-silicon copper tailings (LSCT) using a scanning electron microscope (SEM), and the particle size distribution (PSD) and the activity of LSCT are analysed under the action of mechanical force. Firstly, the current status and application of copper tailings are introduced, and the influence of the particle size of LSCT on its practical application performance is explained. Secondly, the LSCT SEM image target recognition model is designed based on the convolutional neural network (CNN), and the model parameters and the reference CNN are selected. Finally, the experimental process is designed, a SEM image data set of LSCT is prepared, the model is trained through the training set, and the image recognition test is performed on the produced data set. The experimental results show that when the number of iterations of the CNN is 10, the accuracy of model recognition can be guaranteed. After the action of mechanical force, the PSD of LSCT is mainly concentrated at 1 μm~100 μm; that around 1.4 μm~10 μm is the largest, and the PSD of LSCT around 1.4 μm increases with the increase of action time of mechanical force, but the PSD of the LSCT begins to increase when the grinding time exceeds 150 minutes, and the activity of LSCT reaches the maximum (75.545%) at 150 minutes. The average accuracy of SEM image detection of the model is 86.97%, and the model based on DenseNet shows better recognition accuracy than other models. This study provides a reference for analysing the PSD of LSCT.

show abstract

“…The experimental results show that under the complex industrial background, the improved YOLOv3 network achieves strong robust real-time recognition of 0.8 cm thick needles and KR22 bearing machine parts. In order to solve the problems of difficult identification and classification of small-sample industrial mechanical parts, Li et al [ 11 ] established a convolutional neural network model based on the InceptionNet-V3 pre-trained model through transfer learning. Through data expansion, adjusting the learning rate, and optimizing the algorithm, the optimal model was determined, which improved the recognition and classification performance of the algorithm for small-sample industrial machine parts, and the test accuracy rate reached 99.74%.…”

Section: Introductionmentioning

confidence: 99%

A Novel Efficient Convolutional Neural Algorithm for Multi-Category Aliasing Hardware Recognition

Zhang

Liang

et al. 2022

Sensors

View full text Add to dashboard Cite

When performing robotic automatic sorting and assembly operations of multi-category hardware, there are some problems with the existing convolutional neural network visual recognition algorithms, such as large computing power consumption, low recognition efficiency, and a high rate of missed detection and false detection. A novel efficient convolutional neural algorithm for multi-category aliasing hardware recognition is proposed in this paper. On the basis of SSD, the novel algorithm uses Resnet-50 instead of VGG16 as the backbone feature extraction network, and it integrates ECA-Net and Improved Spatial Attention Block (ISAB): two attention mechanisms to improve the ability of learning and extract target features. Then, we pass the weighted features to extra feature layers to build an improved SSD algorithm. At last, in order to compare the performance difference between the novel algorithm and the existing algorithms, three kinds of hardware with different sizes are chosen to constitute an aliasing scene that can simulate an industrial site, and some comparative experiments have been completed finally. The experimental results show that the novel algorithm has an mAP of 98.20% and FPS of 78, which are better than Faster R-CNN, YOLOv4, YOLOXs, EfficientDet-D1, and original SSD in terms of comprehensive performance. The novel algorithm proposed in this paper can improve the efficiency of robotic sorting and assembly of multi-category hardware.

show abstract

Recognition of industrial machine parts based on transfer learning with convolutional neural network

Cited by 4 publications

References 32 publications

A DenseNet‐based feature weighting convolutional network recognition model and its application in industrial part classification

A DenseNet‐based feature weighting convolutional network recognition model and its application in industrial part classification

Image Analysis Technology in the Detection of Particle Size Distribution and the Activity Effect of Low‐Silicon Copper Tailings

A Novel Efficient Convolutional Neural Algorithm for Multi-Category Aliasing Hardware Recognition

Contact Info

Product

Resources

About