Building Detection in High-Resolution Remote Sensing Images by Enhancing Superpixel Segmentation and Classification Using Deep Learning Approaches

被引:7
|
作者
Benchabana, Ayoub [1 ,2 ]
Kholladi, Mohamed-Khireddine [1 ,3 ]
Bensaci, Ramla [4 ]
Khaldi, Belal [4 ]
机构
[1] Univ El Oued, Dept Comp Sci, El Oued 39000, Algeria
[2] Univ El Oued, Lab Operator Theory & EDP Fdn & Applicat, El Oued 39000, Algeria
[3] Univ Constantine 2, MISC Lab Constantine 2, El Khroub 25016, Algeria
[4] Univ Kasdi Merbah Ouargla, Lab Artificial Intelligence & Data Sci, PB 511, Ouargla 30000, Algeria
关键词
arial imagery; building detection; CNN; superpixels segmentation; VAE; NETWORK;
D O I
10.3390/buildings13071649
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
Accurate building detection is a critical task in urban development and digital city mapping. However, current building detection models for high-resolution remote sensing images are still facing challenges due to complex object characteristics and similarities in appearance. To address this issue, this paper proposes a novel algorithm for building detection based on in-depth feature extraction and classification of adaptive superpixel shredding. The proposed approach consists of four main steps: image segmentation into homogeneous superpixels using a modified Simple Linear Iterative Clustering (SLIC), in-depth feature extraction using an variational auto-encoder (VAE) scale on the superpixels for training and testing data collection, identification of four classes (buildings, roads, trees, and shadows) using extracted feature data as input to an Convolutional Neural Network (CNN), and extraction of building shapes through regional growth and morphological operations. The proposed approach offers more stability in identifying buildings with unclear boundaries, eliminating the requirement for extensive prior segmentation. It has been tested on two datasets of high-resolution aerial images from the New Zealand region, demonstrating superior accuracy compared to previous works with an average F1 score of 98.83%. The proposed approach shows potential for fast and accurate urban monitoring and city planning, particularly in urban areas.
引用
收藏
页数:14
相关论文
共 50 条
  • [31] Automatic crack segmentation using deep high-resolution representation learning
    Chen, Hanshen
    Su, Yishun
    He, Wei
    APPLIED OPTICS, 2021, 60 (21) : 6080 - 6090
  • [32] Building Extraction for Urban Infrastructure Mapping Using Deep Neural Networks and High-Resolution Remote Sensing Data
    Sharma, Surendra Kumar
    Sharma, Vivek
    Mishra, Vishal
    JOURNAL OF THE INDIAN SOCIETY OF REMOTE SENSING, 2025,
  • [33] Semi-Supervised Building Detection from High-Resolution Remote Sensing Imagery
    Zheng, Daoyuan
    Kang, Jianing
    Wu, Kaishun
    Feng, Yuting
    Guo, Han
    Zheng, Xiaoyun
    Li, Shengwen
    Fang, Fang
    SUSTAINABILITY, 2023, 15 (15)
  • [34] Boundary-guided DCNN for building extraction from high-resolution remote sensing images
    Yang, Sihan
    He, Qiang
    Lim, Jae Hak
    Jeon, Gwanggil
    INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY, 2024, 132 (9-10) : 5171 - 5171
  • [35] Semantic Segmentation of Urban Remote Sensing Images Based on Deep Learning
    Liu, Jingyi
    Wu, Jiawei
    Xie, Hongfei
    Xiao, Dong
    Ran, Mengying
    APPLIED SCIENCES-BASEL, 2024, 14 (17):
  • [36] RSLC-Deeplab: A Ground Object Classification Method for High-Resolution Remote Sensing Images
    Yu, Zhimin
    Wan, Fang
    Lei, Guangbo
    Xiong, Ying
    Xu, Li
    Ye, Zhiwei
    Liu, Wei
    Zhou, Wen
    Xu, Chengzhi
    ELECTRONICS, 2023, 12 (17)
  • [37] HIGH RESOLUTION REMOTE SENSING IMAGE CLASSIFICATION OF ANCIENT VILLAGES BASED ON DEEP LEARNING
    Chen, Fei
    Fang, Jun
    Hu, Jun
    FRESENIUS ENVIRONMENTAL BULLETIN, 2021, 30 (04): : 3310 - 3324
  • [38] Oil Palm Tree Detection and Health Classification on High-Resolution Imagery Using Deep Learning
    Yarak, Kanitta
    Witayangkurn, Apichon
    Kritiyutanont, Kunnaree
    Arunplod, Chomchanok
    Shibasaki, Ryosuke
    AGRICULTURE-BASEL, 2021, 11 (02): : 1 - 17
  • [39] Vectorized building extraction from high-resolution remote sensing images using spatial cognitive graph convolution model
    Du, Zhuotong
    Sui, Haigang
    Zhou, Qiming
    Zhou, Mingting
    Shi, Weiyue
    Wang, Jianxun
    Liu, Junyi
    ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, 2024, 213 : 53 - 71
  • [40] Deep Learning and Machine Learning for Object Detection in Remote Sensing Images
    Yang, Guowei
    Luo, Qiang
    Yang, Yinding
    Zhuang, Yin
    SIGNAL AND INFORMATION PROCESSING, NETWORKING AND COMPUTERS, 2018, 473 : 249 - 256