{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,2]],"date-time":"2026-05-02T00:47:36Z","timestamp":1777682856255,"version":"3.51.4"},"reference-count":19,"publisher":"SAGE Publications","issue":"3","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JHS"],"published-print":{"date-parts":[[2021,9,23]]},"abstract":"<jats:p>As the need of power supply is tremendously increasing in modern society, the stableness and reliability of the power delivery system are the two essential factors that ensure the power supply safety. With the quick expansion of electricity infrastructures, the failures of power transmission system are becoming more frequent, leading to economic loss and high risk of maintenance work under hazardous conditions. The existing automatic power line inspection utilizes advanced convolutional neural network (CNN) to improve the inspection efficiency, emerging as one promising solution. But the needed computational complexity is high since CNN inference demands large amount of multiplication-and-accumulation operations. In this paper, we alleviate this problem by utilizing the heterogeneous computing techniques to design a real-time on-site inspection system. Firstly, the required computational complexity of CNN inference is reduced using FFT-based convolution algorithms, speeding up the inference. Then we utilize the region of interest (ROI) extrapolation to predict the object detection bounding boxes without CNN inference, thus saving computing power. Finally, a heterogeneous computing architecture is presented to accommodate the requirements of proposed algorithms. According to the experiment results, the proposed design significantly improves the frame rate of CNN-based inspection visual system applied to power line inspection. The processing frame rate is also drastically improved. Moreover, the precision loss is negligible which means our proposed schemes are applicable for real application scenarios.<\/jats:p>","DOI":"10.3233\/jhs-210662","type":"journal-article","created":{"date-parts":[[2021,9,7]],"date-time":"2021-09-07T11:11:24Z","timestamp":1631013084000},"page":"215-224","source":"Crossref","is-referenced-by-count":4,"title":["Real-time on-site inspection system for power transmission based on heterogeneous computing"],"prefix":"10.1177","volume":"27","author":[{"given":"Xiaohong","family":"Yan","sequence":"first","affiliation":[{"name":"Lvliang Power Supply Bureau of Shanxi Power Grid Co., Ltd., Lvliang, Shanxi, 033000, People\u2019s Republic of China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhigang","family":"Zhao","sequence":"additional","affiliation":[{"name":"Lvliang Power Supply Bureau of Shanxi Power Grid Co., Ltd., Lvliang, Shanxi, 033000, People\u2019s Republic of China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yongqiang","family":"Liu","sequence":"additional","affiliation":[{"name":"Lvliang Power Supply Bureau of Shanxi Power Grid Co., Ltd., Lvliang, Shanxi, 033000, People\u2019s Republic of China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"179","reference":[{"key":"10.3233\/JHS-210662_ref2","doi-asserted-by":"crossref","unstructured":"K.\u00a0He, X.\u00a0Zhang, S.\u00a0Ren and J.\u00a0Sun, Deep residual learning for image recognition, in: IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp.\u00a0770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"10.3233\/JHS-210662_ref3","doi-asserted-by":"crossref","unstructured":"Q.\u00a0Huang, D.\u00a0Wang, Z.\u00a0Dong, Y.\u00a0Gao, Y.\u00a0Cai, T.\u00a0Li, B.\u00a0Wu, K.\u00a0Keutzer and J.\u00a0Wawrzynek, Codenet: Efficient deployment of input-adaptive object detection on embedded FPGAs, in: ACM\/SIGDA International Symposium on Field-Programmable Gate Arrays, 2021, pp.\u00a0206\u2013216.","DOI":"10.1145\/3431920.3439295"},{"issue":"1","key":"10.3233\/JHS-210662_ref4","doi-asserted-by":"publisher","first-page":"86","DOI":"10.2478\/s11772-013-0071-0","article-title":"Block-based motion estimation algorithms\u00a0\u2013 A survey","volume":"21","author":"Jakubowski","year":"2013","journal-title":"Opto-Electronics Review"},{"key":"10.3233\/JHS-210662_ref5","doi-asserted-by":"publisher","first-page":"107","DOI":"10.1016\/j.ijepes.2017.12.016","article-title":"Automatic autonomous vision-based power line inspection: A review of current status and the potential role of deep learning","volume":"99","author":"Jenssen","year":"2018","journal-title":"International Journal of Electrical Power & Energy Systems"},{"key":"10.3233\/JHS-210662_ref6","doi-asserted-by":"crossref","unstructured":"T.\u00a0Jia, Y.\u00a0Ju, R.\u00a0Joseph and J.\u00a0Gu, Ncpu: An embedded neural cpu architecture on resource-constrained low power devices for real-time end-to-end performance, in: IEEE\/ACM International Symposium on Microarchitecture (MICRO), 2020, pp.\u00a01097\u20131109.","DOI":"10.1109\/MICRO50266.2020.00091"},{"key":"10.3233\/JHS-210662_ref7","doi-asserted-by":"publisher","first-page":"149999","DOI":"10.1109\/ACCESS.2020.3016213","article-title":"Fault diagnosis of power transmission lines using a UAV-mounted smart inspection system","volume":"8","author":"Kim","year":"2020","journal-title":"IEEE Access"},{"key":"10.3233\/JHS-210662_ref8","doi-asserted-by":"crossref","unstructured":"A.\u00a0Lavin and S.\u00a0Gray, Fast algorithms for convolutional neural networks, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp.\u00a04013\u20134021.","DOI":"10.1109\/CVPR.2016.435"},{"key":"10.3233\/JHS-210662_ref9","doi-asserted-by":"crossref","unstructured":"W.\u00a0Liu, D.\u00a0Anguelov, D.\u00a0Erhan, C.\u00a0Szegedy, S.\u00a0Reed, C.-Y.\u00a0Fu and A.C.\u00a0Berg, Ssd: Single shot multibox detector, in: European Conference on Computer Vision, Springer, 2016, pp.\u00a021\u201337.","DOI":"10.1007\/978-3-319-46448-0_2"},{"issue":"10","key":"10.3233\/JHS-210662_ref10","doi-asserted-by":"publisher","first-page":"2525","DOI":"10.1109\/TPAMI.2018.2858232","article-title":"Thinet: Pruning CNN filters for a thinner net","volume":"41","author":"Luo","year":"2018","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"10.3233\/JHS-210662_ref11","unstructured":"A.\u00a0Paszke, S.\u00a0Gross, S.\u00a0Chintala, G.\u00a0Chanan, E.\u00a0Yang, Z.\u00a0DeVito, Z.\u00a0Lin, A.\u00a0Desmaison, L.\u00a0Antiga and A.\u00a0Lerer, Automatic differentiation in pytorch, in: NIPS-W, 2017."},{"key":"10.3233\/JHS-210662_ref12","doi-asserted-by":"crossref","unstructured":"J.\u00a0Redmon, S.\u00a0Divvala, R.\u00a0Girshick and A.\u00a0Farhadi, You only look once: Unified, real-time object detection, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp.\u00a0779\u2013788.","DOI":"10.1109\/CVPR.2016.91"},{"key":"10.3233\/JHS-210662_ref13","unstructured":"S.\u00a0Ren, K.\u00a0He, R.\u00a0Girshick and J.\u00a0Sun, Faster R-CNN: Towards real-time object detection with region proposal networks, in: Advances in Neural Information Processing Systems, 2015, pp.\u00a091\u201399."},{"key":"10.3233\/JHS-210662_ref15","doi-asserted-by":"crossref","unstructured":"Z.\u00a0Tian, C.\u00a0Shen, H.\u00a0Chen and T.\u00a0He, Fcos: Fully convolutional one-stage object detection, in: IEEE International Conference on Computer Vision, 2019, pp.\u00a09627\u20139636.","DOI":"10.1109\/ICCV.2019.00972"},{"key":"10.3233\/JHS-210662_ref16","unstructured":"C.-J.\u00a0Wu, D.\u00a0Brooks, K.\u00a0Chen, D.\u00a0Chen, S.\u00a0Choudhury, M.\u00a0Dukhan, K.\u00a0Hazelwood, E.\u00a0Isaac, Y.\u00a0Jia, B.\u00a0Jia et al., Machine learning at Facebook: Understanding inference at the edge, in: IEEE International Symposium on High Performance Computer Architecture (HPCA), 2019, pp.\u00a0331\u2013344."},{"key":"10.3233\/JHS-210662_ref17","doi-asserted-by":"publisher","DOI":"10.3390\/s18072258"},{"key":"10.3233\/JHS-210662_ref18","doi-asserted-by":"crossref","unstructured":"C.\u00a0Zhang and V.\u00a0Prasanna, Frequency domain acceleration of convolutional neural networks on CPU-FPGA shared memory system, in: ACM\/SIGDA International Symposium on Field-Programmable Gate Arrays, 2017, pp.\u00a035\u201344.","DOI":"10.1145\/3020078.3021727"},{"key":"10.3233\/JHS-210662_ref20","doi-asserted-by":"publisher","DOI":"10.1109\/IJCNN.2016.7727606"},{"issue":"8","key":"10.3233\/JHS-210662_ref21","doi-asserted-by":"publisher","first-page":"1738","DOI":"10.1109\/JPROC.2019.2918951","article-title":"Edge intelligence: Paving the last mile of artificial intelligence with edge computing","volume":"107","author":"Zhou","year":"2019","journal-title":"Proceedings of the IEEE"},{"key":"10.3233\/JHS-210662_ref22","doi-asserted-by":"crossref","unstructured":"Y.\u00a0Zhu, A.\u00a0Samajdar, M.\u00a0Mattina and P.\u00a0Whatmough, Euphrates: Algorithm-SoC co-design for low-power mobile continuous vision, in: ACM\/IEEE 45th Annual International Symposium on Computer Architecture (ISCA), IEEE, 2018, pp.\u00a0547\u2013560.","DOI":"10.1109\/ISCA.2018.00052"}],"container-title":["Journal of High Speed Networks"],"original-title":[],"link":[{"URL":"https:\/\/content.iospress.com\/download?id=10.3233\/JHS-210662","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T08:44:19Z","timestamp":1777452259000},"score":1,"resource":{"primary":{"URL":"https:\/\/journals.sagepub.com\/doi\/full\/10.3233\/JHS-210662"}},"subtitle":[],"editor":[{"given":"Alireza","family":"Souri","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]},{"given":"Mu-Yen","family":"Chen","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2021,9,23]]},"references-count":19,"journal-issue":{"issue":"3"},"URL":"https:\/\/doi.org\/10.3233\/jhs-210662","relation":{},"ISSN":["1875-8940","0926-6801"],"issn-type":[{"value":"1875-8940","type":"electronic"},{"value":"0926-6801","type":"print"}],"subject":[],"published":{"date-parts":[[2021,9,23]]}}}