3D Convolutional Neural Network for Low-Light Image Sequence Enhancement in SLAM

被引:5
|
作者
Quan, Yizhuo [1 ,2 ]
Fu, Dong [1 ]
Chang, Yuanfei [1 ]
Wang, Chengbo [1 ]
机构
[1] Chinese Acad Sci, Aerosp Informat Reaserch Inst, 9 Dengzhuang South Rd, Beijing 100094, Peoples R China
[2] Univ Chinese Acad Sci, 19 A Yuquan Rd, Beijing 100049, Peoples R China
关键词
visual inertial system; low-light image sequence enhancement; 3D convolutional neural network; DYNAMIC HISTOGRAM EQUALIZATION; LOCALIZATION; VISION; ROBUST;
D O I
10.3390/rs14163985
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Typical visual simultaneous localization and mapping (SLAM) systems rely on front-end odometry for feature extraction and matching to establish the relations between adjacent images. In a low-light environment, the image obtained by a camera is dim and shows scarce information, hindering the extraction of sufficient stable feature points, consequently undermining visual SLAM. Most existing methods focus on low-light enhancement of a single image, neglecting the strong temporal correlation across images in visual SLAM. We propose a method that leverages the temporal information of an input image sequence to enhance the low-light image and employed the enhanced result to improve the feature extraction and matching quality of visual SLAM. Our method trains a three-dimensional convolutional neural network to estimate pixelwise grayscale transformation curves to obtain a low-light enhancement image. Then, the grayscale transformation curves are iteratively applied to obtain the final enhanced result. The training process of the network does not require any paired reference images. We also introduced a spatial consistency loss for the enhanced image to retain the content and texture of the original image. We further integrated our method into VINS-Mono and compared with similar low-light image enhancement methods on the TUM-VI public dataset. The proposed method provides a lower positioning error. The positioning root-mean-squared error of our method is 19.83% lower than that of Zero-DCE++ in low-light environments. Moreover, the proposed network achieves real-time operation, being suitable for integration into a SLAM system.
引用
收藏
页数:21
相关论文
共 50 条
  • [1] LLCNN: A Convolutional Neural Network for Low-light Image Enhancement
    Tao, Li
    Zhu, Chuang
    Xiang, Guoqing
    Li, Yuan
    Jia, Huizhu
    Xie, Xiaodong
    [J]. 2017 IEEE VISUAL COMMUNICATIONS AND IMAGE PROCESSING (VCIP), 2017,
  • [2] Low-Light Image Enhancement Based on Deep Convolutional Neural Network
    Ma Hongqiang
    Ma Shiping
    Xu Yuelei
    Zhu Mingming
    [J]. ACTA OPTICA SINICA, 2019, 39 (02)
  • [3] Low-Light Image Enhancement Based on Deep Convolutional Neural Network
    Ma H.
    Ma S.
    Xu Y.
    Zhu M.
    [J]. Guangxue Xuebao/Acta Optica Sinica, 2019, 39 (02):
  • [4] Low-Light Image Enhancement by Combining Transformer and Convolutional Neural Network
    Yuan, Nianzeng
    Zhao, Xingyun
    Sun, Bangyong
    Han, Wenjia
    Tan, Jiahai
    Duan, Tao
    Gao, Xiaomei
    [J]. MATHEMATICS, 2023, 11 (07)
  • [5] Improvement of Low-Light Image by Convolutional Neural Network
    Kim, Manbae
    [J]. 2019 IEEE 62ND INTERNATIONAL MIDWEST SYMPOSIUM ON CIRCUITS AND SYSTEMS (MWSCAS), 2019, : 189 - 192
  • [6] A Pipeline Neural Network for Low-Light Image Enhancement
    Guo, Yanhui
    Ke, Xue
    Ma, Jie
    Zhang, Jun
    [J]. IEEE ACCESS, 2019, 7 : 13737 - 13744
  • [7] Low-Light Image Restoration Using a Convolutional Neural Network
    Hussain, Syed Ali
    Chalicham, Nandini
    Garine, Likhita
    Chunduru, Shushma
    Nikitha, V.N.V.S.L.
    Prasad V, P.N.S.B.S.V.
    Sanki, Pradyut Kumar
    [J]. Journal of Electronic Materials, 2024, 53 (07) : 3582 - 3593
  • [8] Low-Light Image Restoration Using a Convolutional Neural Network
    Hussain, Syed Ali
    Chalicham, Nandini
    Garine, Likhita
    Chunduru, Shushma
    Nikitha, V. N. V. S. L.
    Prasad, P. N. S. B. S., V
    Sanki, Pradyut Kumar
    [J]. JOURNAL OF ELECTRONIC MATERIALS, 2024,
  • [9] LOW-LIGHT IMAGE ENHANCEMENT USING ASYMMETRIC CONVOLUTIONAL NEURAL NETWORKS
    Liu, Jiajia
    Deng, Zhixiang
    [J]. INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL, 2024, 20 (02): : 479 - 496
  • [10] Low-Light Image Enhancement Based on Multi-Branch All Convolutional Neural Network
    Wu Ruoyou
    Wang Dexing
    Yuan Hongchun
    Peng, Gong
    Chen Guanqi
    Dan, Wang
    [J]. LASER & OPTOELECTRONICS PROGRESS, 2020, 57 (14)