TreeDetector: Using Deep Learning for the Localization and Reconstruction of Urban Trees from High-Resolution Remote Sensing Images

Haoyu Gong,Qian Sun,Chenrong Fang,Le Sun,Ran Su

DOI: https://doi.org/10.3390/rs16030524

IF: 5

2024-01-31

Remote Sensing

Abstract:There have been considerable efforts in generating tree crown maps from satellite images. However, tree localization in urban environments using satellite imagery remains a challenging task. One of the difficulties in complex urban tree detection tasks lies in the segmentation of dense tree crowns. Currently, methods based on semantic segmentation algorithms have made significant progress. We propose to split the tree localization problem into two parts, dense clusters and single trees, and combine the target detection method with a procedural generation method based on planting rules for the complex urban tree detection task, which improves the accuracy of single tree detection. Specifically, we propose a two-stage urban tree localization pipeline that leverages deep learning and planting strategy algorithms along with region discrimination methods. This approach ensures the precise localization of individual trees while also facilitating distribution inference within dense tree canopies. Additionally, our method estimates the radius and height of trees, which provides significant advantages for three-dimensional reconstruction tasks from remote sensing images. We compare our results with other existing methods, achieving an 82.3% accuracy in individual tree localization. This method can be seamlessly integrated with the three-dimensional reconstruction of urban trees. We visualized the three-dimensional reconstruction of urban trees generated by this method, which demonstrates the diversity of tree heights and provides a more realistic solution for tree distribution generation.

environmental sciences,imaging science & photographic technology,remote sensing,geosciences, multidisciplinary

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve

This paper aims to address the issue of tree localization and reconstruction in urban environments using high-resolution remote sensing images. Specifically, the paper proposes a novel two-stage pipeline method (TreeDetector) that combines object detection and a rule-based procedural generation method to improve the accuracy of individual tree detection and achieve 3D reconstruction.

#### Main Issues
1. **Urban Tree Localization**: Locating trees using satellite images in complex urban environments remains a challenge. Particularly in dense canopies, segmentation tasks become more difficult.
2. **Individual Tree Detection**: Existing methods perform poorly in handling dense canopies, leading to reduced accuracy in individual tree detection.

#### Solutions
1. **Combining Object Detection and Semantic Segmentation**: The tree localization problem is divided into two parts—dense clusters and individual trees. By combining object detection methods with a rule-based procedural generation method, the accuracy of individual tree detection is improved.
2. **Two-Stage Pipeline**: Through deep learning and planting strategy algorithms, as well as regional discrimination methods, precise localization of individual trees is ensured, aiding in the distribution inference within dense canopies.
3. **3D Reconstruction**: Estimating the radius and height of trees provides crucial information for 3D reconstruction from remote sensing images. Experimental results show that this method achieves an accuracy of 82.3% in individual tree localization and can be seamlessly integrated into 3D reconstruction.

Through this approach, the paper aims to enhance the accuracy and efficiency of urban tree management, providing valuable references for urban infrastructure safety, urban planning, sustainable development, and natural resource management.