Because of the poor recognition effect of carbon deposit images of automobile engines in the traditional model and the practical application requirements, a classification method for the fine-grained features of carbon deposit images is proposed. For the problem that it is difficult to focus on the nozzle region with large morphological differences in the nozzle carbon image classification task, a carbon image recognition method based on the CA mechanism is proposed. The weights of different channels and spatial locations are adjusted to improve the focusing effect of the model on key features by adding the CA attention mechanism. To address the problem that it is difficult to fully capture the key spatial features in the task of carbon image classification of the piston top, an adaptive carbon image recognition method based on SG-former is proposed, which effectively extracts the fine-grained features in the key areas by automatically assigning the global attention weights and maximizes the translational invariance of the carbon features in the downsampling process through BlurPool pooling. Eventually, the accuracy of carbon degree recognition in the injector nozzle area and piston top area reaches 88.52% and 88.20%, respectively.
ROEA P, THOMPSONS A, PETROUA T, et al. System for determining piston damage based on carbon deposit growth: US20180030890[P]. 2018-02-01.
[2]
WANGJ Q, DINGW Q, DENGJ, et al. Method for detecting and distinguishing early combustion and carbon deposits of engine based on ion current: NL2029619[P]. 2023-06-02.
ZHANGQ, SHUJ, CHENC, et al. Optimization of pneumonia CT classification model using RepVGG and spatial attention features[J]. Frontiers in Medicine, 2023, 10: 1233724.
LIHaoxuan, SUYanqiong. Road defect detection based on improved YOLOv8 [J]. Journal of Test and Measurement Technology, 2024, 38(5): 506-512. (in Chinese)
[10]
HUANGH, YOUZ, CAIH, et al. Fast detection method for prostate cancer cells based on an integrated ResNet50 and YOLOv5 framework[J]. Computer Methods and Programs in Biomedicine, 2022, 226: 107184.
[11]
ATILAÜ, UÇARM, AKYOLK, et al. Plant leaf disease classification using EfficientNet deep learning model[J]. Ecological Informatics, 2021, 61: 101182.
ZUOHaoxuan, HUANGQicheng, YANGJiahao, et al. Improved YOLOv5s-based detection method for crop yellow leaf curl virus disease[J]. Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(S1): 230-238. (in Chinese)
[14]
CHENL, LINS, LUX, et al. Deep neural network based vehicle and pedestrian detection for autonomous driving: a survey[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 22(6): 3234-3246.
[15]
LIUY, ZHANGZ, LIUX, et al. Deep learning-based image classification for online multi-coal and multi-class sorting[J]. Computers & Geosciences, 2021, 157: 104922.
[16]
LIE, ZENGL, ZHOUZ, et al. Edge AI: on-demand accelerating deep neural network inference via edge computing[J]. IEEE Transactions on Wireless Communications, 2019, 19(1): 447-457.
[17]
HUANGQ, CHENY, WANGZ, et al. Research on discrimination method of carbon deposit degree of automobile engine based on deep learning[J]. Computing and Informatics, 2024, 43(1): 126-148.
[18]
ZHUX, CHENGD, ZHANGZ, et al. An empirical study of spatial attention mechanisms in deep networks[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019: 6687-6696.
[19]
HUJ, SHENL, SUNG. Squeeze-and-excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.
[20]
LIX, WANGW, HUX, et al. Selective kernel networks[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019: 510-519.
[21]
MISHRAS K, RAI G, SAHAS, et al. Efficient channel attention based encoder-decoder approach for image captioning in Hindi[J]. ACM Transactions on Asian and Low-Resource Language Information Processing, 2021, 21(3): 49.
[22]
CAOY, XUJ, LINS, et al. GCNet: non-local networks meet squeeze-excitation networks and beyond[C]//2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), 2019: 1971-1980.
[23]
WANGZ, XUJ, LIUL, et al. RANet: ranking attention network for fast video object segmentation[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019: 3977-3986.
ZEGARRAF C, VARGAS-MACHUCAJ, CORONADOA M. Comparison of CNN and CNN-LSTM architectures for tool wear estimation[C]//2021 IEEE Engineering International Research Conference (EIRCON), 2021: 1-4.
[26]
HOUQ, ZHOUD, FENGJ. Coordinate attention for efficient mobile network design[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021: 13708-13717.
[27]
RENS, YANGX, LIUS, et al. SG-former: self-guided transformer with evolving token reallocation[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV), 2023: 5980-5991.
[28]
DINGX, ZHANGX, MAN, et al. RepVGG: making VGG-style ConvNets great again[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021: 13728-13737.
[29]
ENGSTROML, TRANB, TSIPRASD, et al. A rotation and a translation suffice: fooling CNNs with simple transformations[C]//Proceedings of the International Conference on Learning Representations (ICLR), 2019: 1-21.
[30]
AZULAYA, WEISSY. Why do deep convolutional networks generalize so poorly to small image transformations[J]. Journal of Machine Learning Research, 2019, 20(184): 1-25.
[31]
ZHANGR. Making convolutional networks shift-invariant again[C]//International Conference on Machine Learning, PMLR, 2019: 7324-7334.