Energy-Constrained Model Pruning for Efficient In-Orbit Object Detection in Optical Remote Sensing Images

doi:10.1007/978-981-97-1568-8_4

CUG-IR

	Energy-Constrained Model Pruning for Efficient In-Orbit Object Detection in Optical Remote Sensing Images
	Qiu, Shaohua 1,3; Chen, Du 2; Xu, Xinghua 1; Liu, Jia 2
	2024
会议录名称	Communications in Computer and Information Science
卷号	2057 CCIS
页码	34-49
原始文献类型	Conference article (CA)
摘要	Efficient object detection from optical remote sensing (RS) images has always been an important interpretation task for in-orbit RS applications. In recent years, convolutional neural networks have been widely used for object detection with significantly improved detection accuracy. However, the large detection models pose great challenges for the computing, memory and energy supply of resource-constrained in-orbit platforms. In this paper, we propose an efficient in-orbit object detection method with low memory, computation and energy requirements. The proposed method first integrates the compact modules of GhostNet into the detector and further performs the L1-norm based filter pruning to significantly reduce model size and computational complexity. Besides, we propose to use energy as a key metric in filter pruning, and present a novel energy-guided layer-wise pruning rate estimation method so as to achieve energy-efficient object detection. Comprehensive experiments have shown the effectiveness of the proposed method in terms of model size, computational complexity, latency and energy consumption, while maintaining comparable detection accuracy. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
关键词	Complex networks Computational complexity Convolutional neural networks Energy efficiency Energy utilization Object recognition Optical remote sensing Orbits Constrained resources Efficient object detections Filter pruning In-orbit In-orbit object detection Lightweight CNN Objects detection Optical remote sensing Optical remote sensing image Remote sensing images
DOI	10.1007/978-981-97-1568-8_4
语种	英语
ISSN	1865-0929
收录类别	EI
EI入藏号	20241715950888
出版者	Springer Science and Business Media Deutschland GmbH
EISSN	1865-0937
会议名称	7th International Conference on Space Information Network, SINC 2023
EI主题词	Object detection
会议日期	October 12, 2023 - October 13, 2023
会议地点	Wuhan, China
引用统计	被引频次[WOS]：-1 [WOS记录] [WOS相关记录]
文献类型	会议论文
条目标识符	http://ir.cug.edu.cn/handle/2XU834YA/360987
专题	中国地质大学（武汉）
通讯作者	Qiu, Shaohua
作者单位	1.National Key Laboratory of Electromagnetic Energy, Naval University of Engineering, Wuhan; 430033, China 2.School of Computer Science, China University of Geosciences, Wuhan; 430074, China 3.East Lake Laboratory, Wuhan; 430202, China
推荐引用方式 GB/T 7714	Qiu, Shaohua,Chen, Du,Xu, Xinghua,et al. Energy-Constrained Model Pruning for Efficient In-Orbit Object Detection in Optical Remote Sensing Images[C]:Springer Science and Business Media Deutschland GmbH,2024:34-49.