Benchmarking Anchor-Based and Anchor-Free State-of-the-Art Deep Learning Methods for Individual Tree Detection in RGB High-Resolution Images

Mapping the number of mangrove trees in the Guangdong-Hong Kong-Macao Greater Bay Area

Mangroves are vital components of coastal ecosystems. Due to the complex canopy morphology and dense distribution of mangroves, it is challenging to accurately estimate the density based on satellite data. In this study, a density regression-based mangrove mapping network is proposed. The network can capture the multi-scale characteristics of mangroves through the combination of an attention mechanism and a parallel segmentation path, and its performance is better than existing methods. We then apply it to mapping the Greater Bay Area (GBA) the number of mangrove trees. The results show about 2.55 million mangrove trees in the GBA, with an average density of 782 trees per hectare. The tree number of mangroves on the beach is significantly higher than those distributed along the riverbank. This study is the first to achieve mangrove tree count mapping, opening up new prospects for applying satellite-based mangrove monitoring.

A spectral three-dimensional color space model of tree crown health

Protecting the future of forests in the United States and other countries depends in part on our ability to monitor and map forest health conditions in a timely fashion to facilitate management of emerging threats and disturbances over a multitude of spatial scales. Remote sensing data and technologies have contributed to our ability to meet these needs, but existing methods relying on supervised classification are often limited to specific areas by the availability of imagery or training data, as well as model transferability. Scaling up and operationalizing these methods for general broadscale monitoring and mapping may be promoted by using simple models that are easily trained and projected across space and time with widely available imagery. Here, we describe a new model that classifies high resolution (~1 m2) 3-band red, green, blue (RGB) imagery from a single point in time into one of four color classes corresponding to tree crown condition or health: green healthy crowns, red damaged or dying crowns, gray damaged or dead crowns, and shadowed crowns where the condition status is unknown. These Tree Crown Health (TCH) models trained on data from the United States (US) Department of Agriculture, National Agriculture Imagery Program (NAIP), for all 48 States in the contiguous US and spanning years 2012 to 2019, exhibited high measures of model performance and transferability when evaluated using randomly withheld testing data (n = 122 NAIP state x year combinations; median overall accuracy 0.89-0.90; median Kappa 0.85-0.86). We present examples of how TCH models can detect and map individual tree mortality resulting from a variety of nationally significant native and invasive forest insects and diseases in the US. We conclude with discussion of opportunities and challenges for extending and implementing TCH models in support of broadscale monitoring and mapping of forest health.

Deep Learning-Based Instance Segmentation Method of Litchi Canopy from UAV-Acquired Images

Instance segmentation of fruit tree canopies from images acquired by unmanned aerial vehicles (UAVs) is of significance for the precise management of orchards. Although deep learning methods have been widely used in the fields of feature extraction and classification, there are still phenomena of complex data and strong dependence on software performances. This paper proposes a deep learning-based instance segmentation method of litchi trees, which has a simple structure and lower requirements for data form. Considering that deep learning models require a large amount of training data, a labor-friendly semi-auto method for image annotation is introduced. The introduction of this method allows for a significant improvement in the efficiency of data pre-processing. Facing the high requirement of a deep learning method for computing resources, a partition-based method is presented for the segmentation of high-resolution digital orthophoto maps (DOMs). Citrus data is added to the training set to alleviate the lack of diversity of the original litchi dataset. The average precision (AP) is selected to evaluate the metric of the proposed model. The results show that with the help of training with the litchi-citrus datasets, the best AP on the test set reaches 96.25%.

Canopy Volume Extraction of Citrus reticulate Blanco cv. Shatangju Trees Using UAV Image-Based Point Cloud Deep Learning

Automatic acquisition of the canopy volume parameters of the Citrus reticulate Blanco cv. Shatangju tree is of great significance to precision management of the orchard. This research combined the point cloud deep learning algorithm with the volume calculation algorithm to segment the canopy of the Citrus reticulate Blanco cv. Shatangju trees. The 3D (Three-Dimensional) point cloud model of a Citrus reticulate Blanco cv. Shatangju orchard was generated using UAV tilt photogrammetry images. The segmentation effects of three deep learning models, PointNet++, MinkowskiNet and FPConv, on Shatangju trees and the ground were compared. The following three volume algorithms: convex hull by slices, voxel-based method and 3D convex hull were applied to calculate the volume of Shatangju trees. Model accuracy was evaluated using the coefficient of determination (R2) and Root Mean Square Error (RMSE). The results show that the overall accuracy of the MinkowskiNet model (94.57%) is higher than the other two models, which indicates the best segmentation effect. The 3D convex hull algorithm received the highest R2 (0.8215) and the lowest RMSE (0.3186 m3) for the canopy volume calculation, which best reflects the real volume of Citrus reticulate Blanco cv. Shatangju trees. The proposed method is capable of rapid and automatic acquisition for the canopy volume of Citrus reticulate Blanco cv. Shatangju trees.

Semantic Segmentation of Tree-Canopy in Urban Environment with Pixel-Wise Deep Learning

Urban forests are an important part of any city, given that they provide several environmental benefits, such as improving urban drainage, climate regulation, public health, biodiversity, and others. However, tree detection in cities is challenging, given the irregular shape, size, occlusion, and complexity of urban areas. With the advance of environmental technologies, deep learning segmentation mapping methods can map urban forests accurately. We applied a region-based CNN object instance segmentation algorithm for the semantic segmentation of tree canopies in urban environments based on aerial RGB imagery. To the best of our knowledge, no study investigated the performance of deep learning-based methods for segmentation tasks inside the Cerrado biome, specifically for urban tree segmentation. Five state-of-the-art architectures were evaluated, namely: Fully Convolutional Network; U-Net; SegNet; Dynamic Dilated Convolution Network and DeepLabV3+. The experimental analysis showed the effectiveness of these methods reporting results such as pixel accuracy of 96,35%, an average accuracy of 91.25%, F1-score of 91.40%, Kappa of 82.80% and IoU of 73.89%. We also determined the inference time needed per area, and the deep learning methods investigated after the training proved to be suitable to solve this task, providing fast and effective solutions with inference time varying from 0.042 to 0.153 minutes per hectare. We conclude that the semantic segmentation of trees inside urban environments is highly achievable with deep neural networks. This information could be of high importance to decision-making and may contribute to the management of urban systems. It should be also important to mention that the dataset used in this work is available on our website.