Abstract:The global increase in observed forest dieback, characterised by the death of tree foliage, heralds widespread decline in forest ecosystems. This degradation causes significant changes to ecosystem services and functions, including habitat provision and carbon sequestration, which can be difficult to detect using traditional monitoring techniques, highlighting the need for large-scale and high-frequency monitoring. Contemporary developments in the instruments and methods to gather and process data at large-scales mean this monitoring is now possible. In particular, the advancement of low-cost drone technology and deep learning on consumer-level hardware provide new opportunities. Here, we use an approach based on deep learning and vegetation indices to assess crown dieback from RGB aerial data without the need for expensive instrumentation such as LiDAR. We use an iterative approach to match crown footprints predicted by deep learning with field-based inventory data from a Mediterranean ecosystem exhibiting drought-induced dieback, and compare expert field-based crown dieback estimation with vegetation index-based estimates. We obtain high overall segmentation accuracy (mAP: 0.519) without the need for additional technical development of the underlying Mask R-CNN model, underscoring the potential of these approaches for non-expert use and proving their applicability to real-world conservation. We also find colour-coordinate based estimates of dieback correlate well with expert field-based estimation. Substituting ground truth for Mask R-CNN model predictions showed negligible impact on dieback estimates, indicating robustness. Our findings demonstrate the potential of automated data collection and processing, including the application of deep learning, to improve the coverage, speed and cost of forest dieback monitoring.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the large - scale and low - cost monitoring of forest crown dieback. Specifically, the authors aim to develop a deep - learning - based segmentation method to assess the dieback of tree canopies using RGB images taken by drones, without relying on expensive equipment such as LiDAR. ### Background and Problem Description of the Paper The phenomenon of forest crown dieback observed worldwide is increasing, which indicates the widespread decline of forest ecosystems. This degradation will lead to significant changes in ecosystem services and functions, including the loss of habitat provision and carbon sink functions, and these changes are difficult to detect by traditional monitoring techniques. Therefore, there is an urgent need for large - scale and high - frequency monitoring. However, traditional ground - based and drone - based monitoring techniques are not cost - effective when applied on a large scale. To meet this challenge, the authors take advantage of low - cost drone technology and the progress of deep learning on consumer - level hardware, and propose a new method to achieve low - cost and efficient monitoring of crown dieback. ### Specific Problems and Solutions 1. **Requirement for large - scale and high - frequency monitoring**: Traditional monitoring methods are limited in coverage and frequency and cannot capture changes in forest health in a timely manner. To this end, the authors have developed a deep - learning - based segmentation method that can extract the outlines of individual tree crowns from RGB images taken by drones and assess their degree of dieback. 2. **Low - cost and high - efficiency**: This method does not require expensive LiDAR equipment or cumbersome field work, thereby reducing the cost and technical threshold of monitoring, enabling non - expert users to use these tools for effective forest monitoring as well. 3. **Accuracy verification**: The authors verified the accuracy of the method by matching the predicted tree - crown outlines with field - measurement data. They also compared the dieback estimates based on vegetation indices with the results of expert field assessments and found a high correlation between the two, proving the effectiveness of the method. ### Main Contributions - **Automated data collection and processing**: By combining deep learning and drone technology, large - scale, rapid and low - cost monitoring of forest crown dieback has been achieved. - **High accuracy**: Without additional technology development, a relatively high segmentation accuracy (mAP: 0.519) has been achieved and robustness has been maintained in different regions. - **Potential for practical applications**: This method is not only applicable to the research field, but can also provide strong support for actual forest protection and management, especially in dealing with the threats brought by climate change. In conclusion, this paper shows how to use modern technological means to improve forest health monitoring, increasing the coverage, speed and cost - effectiveness of monitoring, and providing new tools and ideas for solving the challenges of large - scale forest protection.

Low-Cost Tree Crown Dieback Estimation Using Deep Learning-Based Segmentation

Individual Tree-Crown Detection in RGB Imagery Using Semi-Supervised Deep Learning Neural Networks

Instance segmentation of deadwood objects in combined optical and elevation data using convolutional neural networks

Deep Learning based Automated Forest Health Diagnosis from Aerial Images

Classification of Single Tree Decay Stages from Combined Airborne LiDAR Data and CIR Imagery

Deep Learning Enables Image-Based Tree Counting, Crown Segmentation, and Height Prediction at National Scale.

Deep Learning in Forestry Using UAV-Acquired RGB Data: A Practical Review

Automated Detection of Conifer Seedlings in Drone Imagery Using Convolutional Neural Networks

Evaluating Different Deep Learning Approaches for Tree Health Classification using High-Resolution Multispectral UAV Data in the Black Forest, Harz Region, and Göttinger Forest

3D-CNN with Multi-Scale Fusion for Tree Crown Segmentation and Species Classification

Precision Detection and Assessment of Ash Death and Decline Caused by the Emerald Ash Borer Using Drones and Deep Learning

Classification of Bark Beetle-Induced Forest Tree Mortality using Deep Learning

Tree Crown Segmentation and Diameter at Breast Height Prediction Based on BlendMask in Unmanned Aerial Vehicle Imagery

A Deep Semantic Segmentation Approach to Map Forest Tree Dieback in Sentinel-2 Data

Unveiling the Potential of Drone-Borne Optical Imagery in Forest Ecology: A Study on the Recognition and Mapping of Two Evergreen Coniferous Species

Individual Tree-Crown Detection and Species Identification in Heterogeneous Forests Using Aerial RGB Imagery and Deep Learning

Comparative Analysis of Machine Learning Techniques and Data Sources for Dead Tree Detection: What Is the Best Way to Go?

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

Individual Tree Detection and Crown Segmentation Based on Metabolic Theory from Airborne Laser Scanning Data

Comparison of Classical Methods and Mask R-CNN for Automatic Tree Detection and Mapping Using UAV Imagery