Rebalanced Region-Pixel Loss for Concrete Defects Segmentation

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 工艺装备与智能机器人研究室

	Rebalanced Region-Pixel Loss for Concrete Defects Segmentation
	Hong KL(洪坤龙)2,3,4; Wang HG(王洪光)3,4; Yang L(杨亮)1; Yuan BB(袁兵兵)3,4
	2021
会议日期	September 17-19, 2021
会议地点	Virtual, Online, China
关键词	Defects Semantic Segmentation Loss Function Concrete Surface Inspection Rebalanced training
页码	137-143
英文摘要	The deep convolution neural network (DCNN) using cross-entropy loss has already achieved high accuracy on the single defect classification task. However, for multi-defect segmentation, the imbalance of categories makes the segmentation of small targets (cracks) not meticulous enough. It also has to adjust the weight of categories manually. This research proposes a region-pixel loss function, which uses a rebalanced training method to balance category weights automatically and classify small categories more accurately. First, we use a wall-climbing robot to obtain the color and depth (RGB-D) information of the surface. Then, we adopt the visual based simultaneous localization and mapping (V-SLAM) method to select key-frame as the input of DCNN. Finally, reconstructing the 3D model of the environment with the output segmentation results. At the same time, to address the weak links of human-computer interaction in the post-processing of defect monitoring, a fast searching method from global 3D model to local details is proposed so that the staff can calmly obtain the local defects' detailed image. Experiments show that our loss function improves the intersection over union (IoU) of small flaws on the verification set. The designed global-to-local interaction method has been applied in the actual dam surface detection and provided an automated solution for monitoring large concrete surfaces.
产权排序	1
会议录	AIPR 2021 - 2021 4th International Conference on Artificial Intelligence and Pattern Recognition
会议录出版者	ACM
会议录出版地	New York
语种	英语
ISBN号	978-1-4503-8408-7
内容类型	会议论文
源URL	[http://ir.sia.cn/handle/173321/30578]
专题	工艺装备与智能机器人研究室
通讯作者	Hong KL(洪坤龙)
作者单位	1.The CCNY Robotics Lab Electrical Engineering Department, City College of New York New York, United States 2.University of Chinese Academy of Sciences, Beijing 100049, China 3.Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110169, China 4.State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China
推荐引用方式 GB/T 7714	Hong KL,Wang HG,Yang L,et al. Rebalanced Region-Pixel Loss for Concrete Defects Segmentation[C]. 见:. Virtual, Online, China. September 17-19, 2021.