Research on YOLOv3 model compression strategy for UAV deployment
UAVs are often limited by limited resources when performing flight tasks, especially the contradiction between storage resources and computing resources when the huge YOLOv3 model is deployed on the edge UAVs. In this paper, we tend to compress YOLOv3 model in different aspects to achieve load avail...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co. Ltd.
2024-01-01
|
| Series: | Cognitive Robotics |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2667241323000381 |
| _version_ | 1797454288770826240 |
|---|---|
| author | Fei Xu Litao Huang Xiaoyang Gao Tingting Yu Leyi Zhang |
| author_facet | Fei Xu Litao Huang Xiaoyang Gao Tingting Yu Leyi Zhang |
| author_sort | Fei Xu |
| collection | DOAJ |
| description | UAVs are often limited by limited resources when performing flight tasks, especially the contradiction between storage resources and computing resources when the huge YOLOv3 model is deployed on the edge UAVs. In this paper, we tend to compress YOLOv3 model in different aspects to achieve load availability at the edge. In this paper, deep separable convolution is introduced to reduce the computation of the model. Then, PR regularization term is used as the regularization term of sparse training to better distinguish scaling factors, and then the hybrid pruning combining channel pruning and layer pruning is carried out on the model according to scaling factors, in order to reduce the number of model parameters and the amount of calculation. Finally, since the training data is a 32-bit floating point number, DoReFa-Net quantization method is used to quantify the model, so as to compress the storage capacity of the model. The experimental results show that the compression scheme proposed in this paper can effectively reduce the number of parameters by 97.5 % and the calculation amount by 82.3 %, and can maintain the original detection efficiency of UAVs. |
| first_indexed | 2024-03-09T15:35:09Z |
| format | Article |
| id | doaj.art-7d110e2297a04b82984b2c797b33fc60 |
| institution | Directory Open Access Journal |
| issn | 2667-2413 |
| language | English |
| last_indexed | 2024-03-09T15:35:09Z |
| publishDate | 2024-01-01 |
| publisher | KeAi Communications Co. Ltd. |
| record_format | Article |
| series | Cognitive Robotics |
| spelling | doaj.art-7d110e2297a04b82984b2c797b33fc602023-11-26T05:14:25ZengKeAi Communications Co. Ltd.Cognitive Robotics2667-24132024-01-014818Research on YOLOv3 model compression strategy for UAV deploymentFei Xu0Litao Huang1Xiaoyang Gao2Tingting Yu3Leyi Zhang4Xi'an University of Technology, Xi'an 710021, Shaanxi, ChinaBeijing Xuanyu Information Technology Co., Ltd., China; Beijing Control Engineering Research Institute, ChinaXi'an University of Technology, Xi'an 710021, Shaanxi, China; Corresponding author.Xi'an University of Technology, Xi'an 710021, Shaanxi, ChinaXi'an University of Technology, Xi'an 710021, Shaanxi, ChinaUAVs are often limited by limited resources when performing flight tasks, especially the contradiction between storage resources and computing resources when the huge YOLOv3 model is deployed on the edge UAVs. In this paper, we tend to compress YOLOv3 model in different aspects to achieve load availability at the edge. In this paper, deep separable convolution is introduced to reduce the computation of the model. Then, PR regularization term is used as the regularization term of sparse training to better distinguish scaling factors, and then the hybrid pruning combining channel pruning and layer pruning is carried out on the model according to scaling factors, in order to reduce the number of model parameters and the amount of calculation. Finally, since the training data is a 32-bit floating point number, DoReFa-Net quantization method is used to quantify the model, so as to compress the storage capacity of the model. The experimental results show that the compression scheme proposed in this paper can effectively reduce the number of parameters by 97.5 % and the calculation amount by 82.3 %, and can maintain the original detection efficiency of UAVs.http://www.sciencedirect.com/science/article/pii/S2667241323000381Deep learningUAVModel compression |
| spellingShingle | Fei Xu Litao Huang Xiaoyang Gao Tingting Yu Leyi Zhang Research on YOLOv3 model compression strategy for UAV deployment Cognitive Robotics Deep learning UAV Model compression |
| title | Research on YOLOv3 model compression strategy for UAV deployment |
| title_full | Research on YOLOv3 model compression strategy for UAV deployment |
| title_fullStr | Research on YOLOv3 model compression strategy for UAV deployment |
| title_full_unstemmed | Research on YOLOv3 model compression strategy for UAV deployment |
| title_short | Research on YOLOv3 model compression strategy for UAV deployment |
| title_sort | research on yolov3 model compression strategy for uav deployment |
| topic | Deep learning UAV Model compression |
| url | http://www.sciencedirect.com/science/article/pii/S2667241323000381 |
| work_keys_str_mv | AT feixu researchonyolov3modelcompressionstrategyforuavdeployment AT litaohuang researchonyolov3modelcompressionstrategyforuavdeployment AT xiaoyanggao researchonyolov3modelcompressionstrategyforuavdeployment AT tingtingyu researchonyolov3modelcompressionstrategyforuavdeployment AT leyizhang researchonyolov3modelcompressionstrategyforuavdeployment |