Search

Rapid and accurate tree-crown detection is significant to forestry management and precision forestry. In the past few decades, the development and maturity of remote sensing technology has created more convenience for tree-crown detection and planting management. However, the variability of the data source leads to significant differences between feature distributions, bringing great challenges for traditional deep-learning-based methods on cross-regional detection. Moreover, compared with other tasks, tree-crown detection has the problems of a poor abundance of objects, an overwhelming number of easy samples and the existence of a quantity of impervious background similar to the tree crown, which make it difficult for the classifier to learn discriminative features. To solve these problems, we apply domain adaptation (DA) to tree-crown detection and propose a DA cascade tree-crown detection framework with multiple region proposal networks, dubbed CAS-DA, realizing cross-regional tree-crown detection and counting from multiple-source remote sensing images. The essence of the multiple region proposal networks in CAS-DA is obtaining the multilevel features and enhancing deeper label classifiers gradually by filtering simple samples of source domain at an early stage. Then, the cascade structure is integrated with a DA object detector and the end-to-end training is realized through the proposed cascade loss function. Moreover, a filtering strategy based on the planting rules of tree crowns is designed and applied to filter wrongly detected trees by CAS-DA. We verify the effectiveness of our method in two different domain shift scenarios, including adaptation between satellite and drone images and cross-satellite adaptation. The results show that, compared to the existing DA methods, our method achieves the best average F1-score in all adaptions. It is also found that the performance between satellite and drone images is significantly worse than that between different satellite images, with average F1-scores of 68.95% and 88.83%, respectively. Nevertheless, there is an improvement of 11.88%~40.00% in the former, which is greater than 0.50%~5.02% in the latter. The above results prove that in tree-crown detection, it is more effective for the DA detector to improve the detection performance on the source domain than to diminish the domain shift alone, especially when a large domain shift exists.

Additional file 1: Supplementary Figure 1. Venn diagram of common and differentially expressed mRNAs of gill tissues of abalone during short-term low salinity exposure: H. discus hannai (DD) and hybrid H. discus hannai ♀ × H. fulgens ♂ (DF). A) Comparisons within DD (CDvsD3, CDvsD24, D3vsD24), B) Comparisons within DF (CFvsF3, CFvsF24, F3vsF24), and C) Comparisons between DD and DF (CDvsCF, D3vsF3, D24vsF24). Controls: CD, CF; 3 h at low salinity exposure: D3, F3; and 24 h at low salinity exposure: D24, F24. Supplementary Fig. 2. Volcano plots of the differentially expressed genes (DEGs) of gill tissues of abalone during short-term low salinity exposure: I) H. discus hannai (DD): A. Comparison in DD between control (CD) and low salinity after 3 h (D3), B. Comparison in DD between control (CD) and low salinity after 24 h (D24). II) Hybrid H. discus hannai ♀ × H. fulgens ♂ (DF). A. Comparison in DF between control (CF) and low salinity after 3 h (F3) B. Comparison in DF between control (CF) and low salinity after 24 h (D24). Red color denotes up-regulated genes and Blue color denotes and down-regulated. Supplementary Fig. 3. Comparison of top 20 Kyoto Encyclopedia of Gene and Genome (KEGG) pathways enrichment statistics of gill tissues of abalone during short-term low salinity exposure. H. discus hannai (DD) and hybrid H.discus hannai ♀ × H. fulgens ♂ (DF). FSS pathway is highlighted with a rectangular red box. (I) Enrichment analysis for DD: A. Comparison between control (CD) and low salinity after 3 h (D3), B. Comparison between control (CD) and low salinity after 24 h (D24), and C. Comparison between low salinity groups after 3 h and 24 h (CD3vsD24). (II) Enrichment analysis for DF: A. Comparison between control (CF) and low salinity group after 3 h (F3), B. Comparison between control (CF) and low salinity group after 24 h (F24), and C. Comparison between low salinity groups after 3 h and 24 h (F3vsF24). The size of each point represents the number of genes annotated to the KEGG pathway. Different colors from yellow to mauve represent the p-value of the enrichment. Supplementary Table S1. Summary of RNA-Seq data quality analysis of H. discus hannai (DD) and hybrid H. discus hannai ♀ × H. fulgens ♂ (DF) during short-term low salinity exposure