Search

Tree canopy height is one of the most important indicators of forest biomass, productivity, and species diversity, but it is challenging to measure accurately from the ground and from space. Here, we used a U-Net model adapted for regression to map the canopy height of all trees in the state of California with very high-resolution aerial imagery (60 cm) from the USDA-NAIP program. The U-Net model was trained using canopy height models computed from aerial LiDAR data as a reference, along with corresponding RGB-NIR NAIP images collected in 2020. We evaluated the performance of the deep-learning model using 42 independent 1 km$^2$ sites across various forest types and landscape variations in California. Our predictions of tree heights exhibited a mean error of 2.9 m and showed relatively low systematic bias across the entire range of tree heights present in California. In 2020, trees taller than 5 m covered ~ 19.3% of California. Our model successfully estimated canopy heights up to 50 m without saturation, outperforming existing canopy height products from global models. The approach we used allowed for the reconstruction of the three-dimensional structure of individual trees as observed from nadir-looking optical airborne imagery, suggesting a relatively robust estimation and mapping capability, even in the presence of image distortion. These findings demonstrate the potential of large-scale mapping and monitoring of tree height, as well as potential biomass estimation, using NAIP imagery., Comment: 29 pages, 9 figures, submitted to Remote Sensing in Ecology and Conservation (RSEC)

Deep learning self-supervised algorithms that can segment an image in a fixed number of hard labels such as the k-means algorithm and relying only on deep learning techniques are still lacking. Here, we introduce the k-textures algorithm which provides self-supervised segmentation of a 4-band image (RGB-NIR) for a $k$ number of classes. An example of its application on high resolution Planet satellite imagery is given. Our algorithm shows that discrete search is feasible using convolutional neural networks (CNN) and gradient descent. The model detects $k$ hard clustering classes represented in the model as $k$ discrete binary masks and their associated $k$ independently generated textures, that combined are a simulation of the original image. The similarity loss is the mean squared error between the features of the original and the simulated image, both extracted from the penultimate convolutional block of Keras 'imagenet' pretrained VGG-16 model and a custom feature extractor made with Planet data. The main advances of the k-textures model are: first, the $k$ discrete binary masks are obtained inside the model using gradient descent. The model allows for the generation of discrete binary masks using a novel method using a hard sigmoid activation function. Second, it provides hard clustering classes -- each pixels has only one class. Finally, in comparison to k-means, where each pixel is considered independently, here, contextual information is also considered and each class is not associated only to similar values in the color channels but also to a texture. Our approach is designed to ease the production of training samples for satellite image segmentation and the k-textures architecture could be adapted to support different number of bands and for more complex tasks, such as object self-segmentation. The model codes and weights are available at https://doi.org/10.5281/zenodo.6359859, Comment: 19 pages, 10 figures, submitted to Frontiers in Environmental Science, section Environmental Informatics and Remote Sensing, Research Topic: Advances in Machine Learning and Deep Learning for Monitoring Terrestrial Ecosystems

Deep learning self supervised algorithms that can segment an image in a fixed number of hard labels such as the k-means algorithm and only relying only on deep learning techniques are still lacking. Here, we introduce the k-textures algorithm which provides self supervised segmentation of a 4-band image (RGB-NIR) for a $k$ number of classes. An example of its application on high resolution Planet satellite imagery is given. Our algorithm shows that discrete search is feasible using convolutional neural networks (CNN) and gradient descent. The model detects $k$ hard clustering classes represented in the model as $k$ discrete binary masks and their associated $k$ independently generated textures, that combined are a simulation of the original image. The similarity loss is the mean squared error between the features of the original and the simulated image, both extracted from the penultimate convolutional block of Keras 'imagenet' pretrained VGG-16 model and a custom feature extractor made with Planet data. The main advances of the k-textures model are: first, the $k$ discrete binary masks are obtained inside the model using gradient descent. The model allows for the generation of discrete binary masks using a novel method using a hard sigmoid activation function. Second, it provides hard clustering classes -- each pixels has only one class. Finally, in comparison to k-means, where each pixel is considered independently, here, contextual information is also considered and each class is not associated only to a similar values in the color channels but to a texture. Our approach is designed to ease the production of training samples for satellite image segmentation. The model codes and weights are available at https://doi.org/10.5281/zenodo.6359859, 19 pages, 10 figures, submitted to Frontiers in Environmental Science, section Environmental Informatics and Remote Sensing, Research Topic: Advances in Machine Learning and Deep Learning for Monitoring Terrestrial Ecosystems

This pdf contains all supplementary figures, tables, and texts cited in the main manuscript.

Supplementary Tables S1-S3 and Figures S1-S12

1. Canopy gaps are openings in the forest canopy resulting from branch fall and tree mortality events. The geographical distribution of large canopy gaps may reflect underlying variation in mortality and growth processes. However, a lack of data at the appropriate scale has limited our ability to study this relationship until now. 2. We detected canopy gaps using a unique LiDAR data set consisting of 650 transects randomly distributed across 2500 km2 of the Brazilian Amazon. We characterized the size distribution of canopy gaps using a power-law and we explore the variation in the exponent, α. We evaluated how the α varies across the Amazon, in response to disturbance by humans and natural environmental processes that influence tree mortality rates. 3. We observed that South-eastern forests contained a higher proportion of large gaps than North-western, which is consistent with recent work showing greater tree mortality rates in the Southeast than the Northwest. Regions characterised by strong wind gust speeds, frequent lightning and greater water shortage also had a high proportion of large gaps, indicating that geographical variation in α is a reflection of underlying disturbance processes. Forests on fertile soils were also found to contain a high proportion of large gaps, in part because trees grow tall on these sites and create large gaps when they fall; thus canopy gap analysis picked up differences in growth as well as mortality processes. Finally, we found that human modified forests had a higher proportion of large gaps than intact forests, as we would expect given that these forests have been disturbed. 4. Synthesis: The proportion of large gaps in the forest canopy varied substantially over the Brazilian Amazon. We have shown that the trends can be explained by geographic variation in disturbance and growth. The frequency of extreme weather events is predicted to increase under climate change, and changes could lead to greater forest disturbance, which should be detectable as an increased proportion of large gaps in intact forests., NE/S010750/1