Neural Radiance Field Dynamic Scene SLAM Based on Ray Segmentation and Bundle Adjustment

This is an early access version, the complete PDF, HTML, and XML versions will be available soon. Open AccessArticle by Yuquan ZhangYuquan Zhang Yuquan Zhang received the B.S. degree in Transportation from Yanshan University, Qinhuangdao,China, Yuquan Zhang received the B.S. degree in Transportation from Yanshan University, Qinhuangdao,China, in 2006, and the M.S. degree in Mechatronics Engineering from Hebei University of Science and Technology, Shijiazhuang, China, in 2014. He is currently pursuing a Ph.D. in Transportation Engineering at Shijiazhuang Tiedao University. His research interests include simultaneous localization and mapping (SLAM), sensor fusion, path planning and related areas. 1,2 and Guosheng FengGuosheng Feng Guosheng Feng is a professor and doctoral supervisor at Shijiazhuang Tiedao University. He graduated Guosheng Feng is a professor and doctoral supervisor at Shijiazhuang Tiedao University. He graduated from Hebei University of Technology with a Bachelor of Engineering degree in Automotive Engineering in 1983 and a Master of Engineering degree in Construction Machinery in 1993. In 2004, he earned his Ph.D. in Power Machinery and Engineering from the Beijing Institute of Technology. From June 2005 to June 2007, he was a postdoctoral fellow at Tianjin University and the China Automotive Technology and Research Center. He was promoted to associate professor in 2001 and exceptionally promoted to professor in 2004. From February 2011 to February 2012, he was a state-sponsored visiting scholar at the University of North Carolina, USA. He has led multiple projects funded by the National Natural Science Foundation of China and major scientific research achievement transformation projects in Hebei Province. He has published dozens of papers indexed by SCI and EI.The main research areas include finite element analysis, DSP, and automotive electronic control technology. 1,* 1 School of Traffic and Transportation, Shijiazhuang Tiedao University, Shijiazhuang 050043, China 2 Department of Automotive Engineering, Hebei Jiaotong Vocational and Tecenical College, Shijiazhuang 050035, China * Author to whom correspondence should be addressed. Sensors 2025, 25(6), 1679; https://doi.org/10.3390/s25061679 (registering DOI) Submission received: 25 January 2025 / Revised: 5 March 2025 / Accepted: 6 March 2025 / Published: 8 March 2025 Abstract The current neural implicit SLAM methods have demonstrated excellent performance in reconstructing ideal static 3D scenes. However, it remains a significant challenge for these methods to handle real scenes with drastic changes in lighting conditions and dynamic environments. This paper proposes a neural implicit SLAM method that effectively deals with dynamic scenes. We employ a keyframe selection and tracking switching approach based on Lucas–Kanade (LK) optical flow, which serves as prior construction for the Conditional Random Fields potential function. This forms a semantic-based joint estimation method for dynamic and static pixels and constructs corresponding loss functions to impose constraints on dynamic scenes. We conduct experiments on various dynamic and challenging scene datasets, including TUM RGB-D, Openloris, and Bonn. The results demonstrate that our method significantly outperforms existing neural implicit SLAM systems in terms of reconstruction quality and tracking accuracy. Share and Cite MDPI and ACS Style Zhang, Y.; Feng, G. Neural Radiance Field Dynamic Scene SLAM Based on Ray Segmentation and Bundle Adjustment. Sensors 2025, 25, 1679. https://doi.org/10.3390/s25061679 AMA Style Zhang Y, Feng G. Neural Radiance Field Dynamic Scene SLAM Based on Ray Segmentation and Bundle Adjustment. Sensors. 2025; 25(6):1679. https://doi.org/10.3390/s25061679 Chicago/Turabian Style Zhang, Yuquan, and Guosheng Feng. 2025. "Neural Radiance Field Dynamic Scene SLAM Based on Ray Segmentation and Bundle Adjustment" Sensors 25, no. 6: 1679. https://doi.org/10.3390/s25061679 APA Style Zhang, Y., & Feng, G. (2025). Neural Radiance Field Dynamic Scene SLAM Based on Ray Segmentation and Bundle Adjustment. Sensors, 25(6), 1679. https://doi.org/10.3390/s25061679 Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here. Article Metrics Article metric data becomes available approximately 24 hours after publication online. Source link Yuquan Zhang www.mdpi.com

Abstract

Share and Cite

Article Metrics

Related Posts

‘We’re losing our environmental history’: The future of government information under Trump

A Novel Model for Optimizing Roundabout Merging Decisions Based on Markov Decision Process and Force-Based Reward Function

Seasonal Variation in Nutritional Substances in Varieties of Leafy Chinese Kale (Brassica oleracea var. alboglabra): A Pilot Trial

Greenberg