A Deep Fusion uNet for Mapping Forests at Tree Species Levels with Multi-Temporal High Spatial Resolution Satellite Imagery

Abstract

It is critical to acquire the information of forest type at the tree species level due to its strong links with various quantitative and qualitative indicators in forest inventories. The efficiency of deep-learning classification models for high spatial resolution (HSR) remote sensing image has been demonstrated with the ongoing development of artificial intelligence technology. However, due to limited statistical separability and complicated circumstances, completely automatic and highly accurate forest type mapping at the tree species level remains a challenge. To deal with the problem, a novel deep fusion uNet model was developed to improve the performance of forest classification refined at the dominant tree species level by combining the beneficial phenological characteristics of the multi-temporal imagery and the powerful features of the deep uNet model. The proposed model was built on a two-branch deep fusion architecture with the deep Res-uNet model functioning as its backbone. Quantitative assessments of China’s Gaofen-2 (GF-2) HSR satellite data revealed that the suggested model delivered a competitive performance in the Wangyedian forest farm, with an overall classification accuracy (OA) of 93.30% and a Kappa coefficient of 0.9229. The studies also yielded good results in the mapping of plantation species such as the Chinese pine and the Larix principis.

Urban Tree Canopy Mapping Based on Double-Branch Convolutional Neural Network and Multi-Temporal High Spatial Resolution Satellite Imagery

“…In the research on using a double-branch convolutional neural network to process satellite images, in addition to most of the double-branch network structures based on the decision fusion strategy, some double-branch convolutional neural network structures using the feature fusion strategy have also been constructed. Existing research has shown that the double-branch convolutional neural network based on the feature fusion strategy has outstanding performance in processing multi-source satellite images [68]. However, no study has combined double-branch convolutional neural networks and multi-temporal satellite imagery as special multi-source data for urban tree canopy mapping.…”

“…With the development of deep learning, multi-branch networks that can effectively process multi-source data with different structures have been constructed and have shown excellent performance [66][67][68][69]. Multi-branch convolutional neural networks with different structures were constructed based on different data fusion strategies, including multibranch neural networks based on a feature fusion strategy and multi-branch networks based on a decision fusion strategy [70][71][72][73][74].…”

Machine Learning-Based Forest Type Mapping from Multi-Temporal Remote Sensing Data: Performance and Comparative Analysis

“…The RF classifier employed achieved high accuracy in forest/non-forest mapping and forest type classification (98.3% and 94.8%, respectively), while the inclusion of topographic data improved the accuracy in identifying eight tree species from 75.6% to 81.7% (approach 1) and up to 89.5% for broadleaf and 82% for coniferous species (approach 2). Guo et al (2021) [11] developed a novel deep fusion uNet model which utilizes both multi-temporal imagery and the deep uNet model's features. The model demonstrated competitive performance in forest classification, achieving an overall accuracy of 93.30% and a Kappa coefficient of 0.9229 for China's Gaofen-2 satellite data.…”

Semantic Segmentation of Remote Sensing Imagery Using an Enhanced Encoder-Decoder Architecture

“…UNet is a CNN that uses an encoder-decoder structure with skip connections to preserve spatial information and reduce the loss of spatial resolution during feature extraction. UNet has been used for various remote sensing applications, including crop monitoring (Fan et al, 2022), forest mapping (Guo et al, 2021), building detection (Liu et al, 2020), road segmentation (Al Saad et al, 2020b), and oil spill detection (El Rai et al, 2020). An advanced version of UNet called RUNet was devised by (Hu et al, 2019).…”