Interoperability Study of Data Preprocessing for Deep Learning and High-Resolution Aerial Photographs for Forest and Vegetation Type Identification

Abstract

When original aerial photographs are combined with deep learning to classify forest vegetation cover, these photographs are often hindered by the interlaced composition of complex backgrounds and vegetation types as well as the influence of different deep learning calculation processes, resulting in unpredictable training and test results. The purpose of this research is to evaluate (1) data preprocessing, (2) the number of classification targets, and (3) convolutional neural network (CNN) approaches combined with deep learning’s effects on high-resolution aerial photographs to identify forest and vegetation types. Data preprocessing is mainly composed of principal component analysis and content simplification (noise elimination). The number of classification targets is divided into 14 types of forest vegetation that are more complex and difficult to distinguish and seven types of forest vegetation that are simpler. We used CNN approaches to compare three CNN architectures: VGG19, ResNet50, and SegNet. This study found that the models had the best execution efficiency and classification accuracy after data preprocessing using principal component analysis. However, an increase in the number of classification targets significantly reduced the classification accuracy. The algorithm analysis showed that VGG19 achieved the best classification accuracy, but SegNet achieved the best performance and overall stability of relative convergence. This proves that data preprocessing helps identify forest and plant categories in aerial photographs with complex backgrounds. If combined with the appropriate CNN algorithm, these architectures will have great potential to replace high-cost on-site forestland surveys. At the end of this study, a user-friendly classification system for practical application is proposed, and its testing showed good results.

Ecological restoration success on the Loess Plateau of China: A qualitative and quantitative exploration based on rephotography

“…In addition, there is no way to validate the satellite imagery‐derived NDVI based on images from the past. Moreover, the resolution of satellite‐based remote sensing images is not high enough to distinguish plant species and to reflect the vegetation succession (Li et al, 2014; Lin & Chuang, 2021; Xie, Sha, & Yu, 2008). For example, little change in NDVI (0.65 in 1998 to 0.71 in 2016) was found at Yangjuangou (Figure 10d and Figure 7i), despite the vegetation succession (from grassland to shrubland) occurring during this period (Wang et al, 2011).…”

Machine learning assisted remote forestry health assessment: a comprehensive state of the art review

“…Likewise, DL models are used in conjunction with RGB, multispectral, and hyper-spectral images, to perform different tasks concerning the assessment of forest health. Lin and Chuang (2021) used deep convolutional neural networks ResNet50, VGG19, and SegNet to extract features from aerial RGB pictures to perform tree classification. However the initial results showed poor performance based on accuracy; thus the authors proposed a simplification of the images using Principal Component Analysis, selecting only the most important features of the images.…”

“…The use of high-resolution cameras has allowed researchers to couple them with deep convolutional neural networks (Osco et al, 2021). Using deep learning structures alongside high-resolution aerial images has had good results in individual tree crown segmentation (Lin and Chuang, 2021;Onishi and Ise, 2021;. Other applications of deep convolutional neural networks are tree identification from aerial RGB and multispectral images, using temporal information has also been explored by researchers with the aid of recurrent convolutional neural networks (Feng et al, 2020).…”

“…Other applications of deep convolutional neural networks are tree identification from aerial RGB and multispectral images, using temporal information has also been explored by researchers with the aid of recurrent convolutional neural networks (Feng et al, 2020). The most common deep learning back-bones used to perform feature extraction are, VGG19, RES-NET and Seg-Net (Pulido et al, 2020;Lin and Chuang, 2021;Hao Z. et al, 2022). Other structures used in semantic segmentation processes are U-NET and Mask-RCNN (Pulido et al, 2020).…”

Evaluation of Decision Fusions for Classifying Karst Wetland Vegetation Using One-Class and Multi-Class CNN Models with High-Resolution UAV Images

“…In contrast, deep learning allows end-to-end learning without recognizing human intervention in its training process, and its deeper architecture of higher complexity allows it to learn more complex features autonomously [22], avoiding complex feature engineering. Among the existing deep learning algorithms, convolutional neural networks (CNNs) have been applied to the classification of land cover types [23][24][25]. Fu et al [26] used DeepLabV3+ and PSPNet algorithms to classify mangrove communities and both achieved over 86% overall accuracy.…”