Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

Abstract

Land cover mapping from satellite images has progressed from visual and statistical approaches to Random Forests (RFs) and, more recently, advanced image recognition techniques such as convolutional neural networks (CNNs). CNNs have a conceptual benefit over RFs in recognising spatial feature context, but potentially at the cost of reduced spatial detail. We tested the use of CNNs for improved land cover mapping based on Landsat data, compared with RFs, for a study area of approximately 500 km × 500 km in southeastern Australia. Landsat 8 geomedian composite surface reflectances were available for 2018. Label data were a simple nine-member land cover classification derived from reference land use mapping (Catchment Scale Land Use of Australia—CLUM), and further enhanced by using custom forest extent mapping (Forests of Australia). Experiments were undertaken testing U-Net CNN for segmentation of Landsat 8 geomedian imagery to determine the optimal combination of input Landsat 8 bands. The results were compared with those from a simple autoencoder as well as an RF model. Segmentation test results for the best performing U-Net CNN models produced an overall accuracy of 79% and weighted-mean F1 score of 77% (9 band input) or 76% (6 band input) for a simple nine-member land cover classification, compared with 73% and 68% (6 band input), respectively, for the best RF model. We conclude that U-Net CNN models can generate annual land cover maps with good accuracy from proxy training data, and can also be used for quality control or improvement of existing land cover products.

Proxying economic activity with daytime satellite imagery: Filling data gaps across time and space

“…Although the application of CNNs to land cover classification could have higher computing-power demands and may require additional region-specific calibration of ground-truth data [e.g. mentioned in ( 60 , 61 )], these additional challenges could be solved by introducing additional instruments such as hyperparameter tuning and model pre-training. Therefore, using CNNs to classify land cover has large potential for future research to investigate regional or even subregional economic activity at a global scale.…”

Towards Improving Sustainable Water Management in Geothermal Fields: SVM and RF Land Use Monitoring

“…By 2023, the importance of Band 4 in chlorophyll absorption remained critical for vegetation health identification; Band 2 was pivotal in providing textural details due to its moisture and shadow sensitivity, and Band 1 was essential for atmospheric corrections to improve forest health assessments. This expanding understanding is vital for establishing adaptive forest management techniques that respond to changing climatic circumstances and enable sustainable water resource management in geothermal fields, thus improving conservation efforts and maintaining ecological balance [101,102].…”

Improvements in Forest Segmentation Accuracy Using a New Deep Learning Architecture and Data Augmentation Technique

“…Meanwhile, U-Net [21] was based on the encoder-decoder structure and was proposed to handle medical image segmentation tasks. It is the most widely used deep learning architecture for remote sensing image segmentation and has achieved satisfactory results [30]. In order to better extract the feature information of NIR band of satellite imagery, inspired by U-Net, we proposed a high-resolution forest remote sensing image segmentation network by fusing multi-scale features based on double input.…”

“…Wei et al [29] demonstrated that the CNN-based image segmentation network achieved much better results than the spectral vegetation index-based method and the machine-learning-based method in the task of mapping the large plateau forest of Sanjiangyuan. Among the prevalent deep learning image segmentation networks, U-Net, which is designed based on the encoder-decoder structure, is the most commonly used deep learning architecture to perform remote sensing image segmentation [30]. In the field of forest remote sensing image segmentation, Freudenberg et al [31] developed a novel method for detecting oil palm plantations using very high-resolution satellite imagery, based on the U-Net architecture.…”