Review and Evaluation of Deep Learning Architectures for Efficient Land Cover Mapping with UAS Hyper-Spatial Imagery: A Case Study Over a Wetland

Abstract

Deep learning has already been proved as a powerful state-of-the-art technique for many image understanding tasks in computer vision and other applications including remote sensing (RS) image analysis. Unmanned aircraft systems (UASs) offer a viable and economical alternative to a conventional sensor and platform for acquiring high spatial and high temporal resolution data with high operational flexibility. Coastal wetlands are among some of the most challenging and complex ecosystems for land cover prediction and mapping tasks because land cover targets often show high intra-class and low inter-class variances. In recent years, several deep convolutional neural network (CNN) architectures have been proposed for pixel-wise image labeling, commonly called semantic image segmentation. In this paper, some of the more recent deep CNN architectures proposed for semantic image segmentation are reviewed, and each model’s training efficiency and classification performance are evaluated by training it on a limited labeled image set. Training samples are provided using the hyper-spatial resolution UAS imagery over a wetland area and the required ground truth images are prepared by manual image labeling. Experimental results demonstrate that deep CNNs have a great potential for accurate land cover prediction task using UAS hyper-spatial resolution images. Some simple deep learning architectures perform comparable or even better than complex and very deep architectures with remarkably fewer training epochs. This performance is especially valuable when limited training samples are available, which is a common case in most RS applications.

A deep‐learning model to predict thunderstorms within 400 km2 South Texas domains

“…Thanks to the increasing availability of large data sets and affordable computational power, deep‐learning algorithms can now model complex nonlinear relationships in the earth sciences (Hernández et al ., 2016; Scher, 2018; Gagne II et al ., 2019; Lagerquist et al ., 2019; Kamangir et al ., 2018; Pashaei et al ., 2020; Reichstein et al ., 2019). The introduction of DLNNs in 2006 (Hinton et al ., 2006) has led to large changes in artificial intelligence (AI) and ML in many areas of research.…”

Research challenges and opportunities for using big data in global change biology

“…Machine learning, particularly deep learning, has been increasingly used to discover new patterns from data. For example, machine learning has benefited the areas of plant specialized metabolism (Moore et al., 2019), plant–pathogen interactions (Sperschneider, 2019), biomass estimation (Anderson et al., 2018; Zhao, Popescu, Meng, Pang, & Agca, 2011), wetland mapping (Pashaei, Kamangir, Starek, & Tissot, 2020), and geoscience (Reichstein et al., 2019). The challenges and opportunities of applying machine learning in those research areas offer important lessons for global change biology.…”

“…, biomass estimation (Anderson et al, 2018; Zhao, Popescu, Meng, Pang, & Agca, 2011), wetland mapping(Pashaei, Kamangir, Starek, & Tissot, 2020), and geoscience(Reichstein et al, 2019). The challenges and opportunities of applying machine learning in those re-search areas offer important lessons for global change biology.…”

NAS-HRIS: Automatic Design and Architecture Search of Neural Network for Semantic Segmentation in Remote Sensing Images

“…Although the above methods based on deep learning greatly enhance the accuracy and efficiency of remote sensing image segmentation [ 24 , 25 , 26 ], a robust model usually requires relevant experts to spend a lot of time and energy to complete it. Feature extraction and fusion are key for robust and effective image processing in remote sensing [ 27 ].…”