Accurate measurements of the structural characteristics of trees such as height, diameter, sweep and taper are an important part of forest inventories in managed forests and commercial plantations. Both terrestrial and aerial LiDAR are currently employed to produce pointcloud data from which inventory metrics can be determined. Terrestrial/ground-based scanning typically provides pointclouds resolutions of many thousands of points per m ${}^2$ from which tree stems can be observed and inventory measurements made directly, whereas typical resolutions from aerial scanning (tens of points per m ${}^2$ ) require inventory metrics to be regressed from LiDAR variables using inventory reference data collected from the ground. Recent developments in miniaturised LiDAR sensors are enabling aerial capture of pointclouds from low-flying aircraft at high-resolutions (hundreds of points per m ${}^2$ ) from which tree stem information starts to become directly visible, enabling the possibility for plot-scale inventories that do not require access to the ground. In this paper, we develop new approaches to automated tree detection, segmentation and stem reconstruction using algorithms based on deep supervised machine learning which are designed for use with aerially acquired high-resolution LiDAR pointclouds. Our approach is able to isolate individual trees, determine tree stem points and further build a segmented model of the main tree stem that encompasses tree height, diameter, taper, and sweep. Through the use of deep learning models, our approach is able to adapt to variations in pointcloud densities and partial occlusions that are particularly prevalent when data is captured from the air. We present results of our algorithms using high-resolution LiDAR pointclouds captured from a helicopter over two Radiata pine forests in NSW, Australia.

Convolutional neural networks (CNNs) have been shown to be a powerful tool for image classification. Recently, they have been adopted into the remote sensing community with applications in material classification from hyperspectral images. However, CNNs are time-consuming to train and often require large amounts of labeled training data. The widespread use of CNNs in the image processing and computer vision communities has been facilitated by the networks that have already been trained on large amounts of data. These pretrained networks can be used to initialize networks for new tasks. This transfer of knowledge makes it far less time-consuming to train a new classifier and reduces the need for a large labeled data set. This concept of transfer learning has not yet been fully explored by those using CNNs to train material classifiers from hyperspectral data. This paper provides an insight into training hyperspectral CNN classifiers by transferring knowledge from well labeled data sets to data sets that are less well labeled. It is shown that these CNNs can transfer between completely different domains and sensing platforms, and still improve classification performance. The application of this work is in the training of material classifiers of data acquired from field-based platforms, by transferring knowledge from publicly accessible airborne data sets. Factors, such as training set size, CNN architectures, and the impact of filter width and wavelength interval, are studied.

This paper proposes novel autoencoders for unsupervised feature-learning from hyperspectral data. Hyperspectral data typically have many dimensions and a significant amount of variability such that many data points are required to represent the distribution of the data. This poses challenges for higher-level algorithms which use the hyperspectral data (e.g., those that map the environment). Feature-learning mitigates this by projecting the data into a lower-dimensional space where the important information is either preserved or enhanced. In many applications, the amount of labelled hyperspectral data that can be acquired is limited. Hence, there is a need for feature-learning algorithms to be unsupervised. This work proposes unsupervised techniques that incorporate spectral measures from the remote-sensing literature into the objective functions of autoencoder feature learners. The proposed techniques are evaluated on the separability of their feature spaces as well as on their application as features for a clustering task, where they are compared against other unsupervised feature-learning approaches on several different datasets. The results show that autoencoders using spectral measures outperform those using the standard squared-error objective function for unsupervised hyperspectral feature-learning.

This paper proposes the Relit Spectral AngleStacked Autoencoder, a novel unsupervised feature learning approach for mapping pixel reflectances to illumination invariant encodings. This work extends the Spectral Angle-Stacked Autoencoder so that it can learn a shadow-invariant mapping. The method is inspired by a deep learning technique, Denoising Autoencoders, with the incorporation of a physics-based model for illumination such that the algorithm learns a shadow invariant mapping without the need for any labelled training data, additional sensors, a priori knowledge of the scene or the assumption of Planckian illumination. The method is evaluated using datasets captured from several different cameras, with experiments to demonstrate the illumination invariance of the features and how they can be used practically to improve the performance of high-level perception algorithms that operate on images acquired outdoors.

Surveying of woody debris left over from harvesting operations on managed forests is an important step in monitoring site quality, managing the extraction of residues and reconciling differences in pre-harvest inventories and actual timber yields. Traditional methods for post-harvest survey involving manual assessment of debris on the ground over small sample plots are labor-intensive, time-consuming, and do not scale well to heterogeneous landscapes. In this paper, we propose and evaluate new automated methods for the collection and interpretation of high-resolution, Unmanned Aerial Vehicle (UAV)-borne imagery over post-harvested forests for estimating quantities of fine and coarse woody debris. Using high-resolution, geo-registered color mosaics generated from UAV-borne images, we develop manual and automated processing methods for detecting, segmenting and counting both fine and coarse woody debris, including tree stumps, exploiting state-of-the-art machine learning and image processing techniques. Results are presented using imagery over a post-harvested compartment in a Pinus radiata plantation and demonstrate the capacity for both manual image annotations and automated image processing to accurately detect and quantify coarse woody debris and stumps left over after harvest, providing a cost-effective and scalable survey method for forest managers.

A machine learning methodology is developed for the detection of individual trees, classification of health, and detection of dead/dying trees in 125 mm resolution aerial multispectral orthoimagery and photogrammetric pointcloud data. The novelty of the proposed method lies in its flexible utilization of features from different georegistered data sources. The methodology is evaluated on a commercial Pinus radiata plantation with a known outbreak of Sirex noctilio. An analysis was carried out to determine the value of the different sensors/data sources and the influence of the spatial resolution of the data on performance. Using the proposed methodology, trees were detected with approximate commission and omission errors of 5% and 22%, respectively, whilst dead or dying trees could be detected with commission and omission errors of approximately 26% and 9%. It was found that the multispectral imagery was the most informative source of data for the given tasks. Tree detection in general was found to be sensitive to the spatial resolution of the data, whilst diseased tree detection was found to be more robust.