A multi-scale dataset with comprehensive analysis for building detection from remote sensing images

The role of object detection in remote sensing images is crucial in urban management and planning. However, there are some aspects that pose serious challenges for the inspection of buildings. For example, remote sensing images have different resolutions and different object sizes. And the object de...

Full description

Saved in:
Bibliographic Details
Published in2023 16th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI) pp. 1 - 5
Main Authors Zhang, Han, Bi, Meizhen, Xu, Tao, Zang, Junyuan
Format Conference Proceeding
LanguageEnglish
Published IEEE 28.10.2023
Subjects
Building detection
Buildings
Image resolution
Machine learning algorithms
Multi-scale
Object detection
remote sensing
Robustness
Signal processing
Signal processing algorithms
Online AccessGet full text

Cover

More Information
Summary:The role of object detection in remote sensing images is crucial in urban management and planning. However, there are some aspects that pose serious challenges for the inspection of buildings. For example, remote sensing images have different resolutions and different object sizes. And the object detection algorithm based on convolutional neural network belongs to supervised machine learning algorithms, which requires the support of large-scale datasets. To overcome these issues, a multi-scale building dataset is proposed in this paper. The dataset comprises over 1200 images derived from 0.8-8 meter remote sensing image data. In the experiments, object detection models such as Faster R-CNN and RTMDet were used to validate the dataset and achieve the expected results. The experiments demonstrate that the dataset enhances the precision of building detection models in remote sensing images with multi-resolution.
DOI:10.1109/CISP-BMEI60920.2023.10373371