Assessment of Deep Learning Techniques for Land Use Land Cover Classification in Southern New Caledonia

Abstract

Land use (LU) and land cover (LC) are two complementary pieces of cartographic information used for urban planning and environmental monitoring. In the context of New Caledonia, a biodiversity hotspot, the availability of up-to-date LULC maps is essential to monitor the impact of extreme events such as cyclones and human activities on the environment. With the democratization of satellite data and the development of high-performance deep learning techniques, it is possible to create these data automatically. This work aims at determining the best current deep learning configuration (pixel-wise vs semantic labelling architectures, data augmentation, image prepossessing, …), to perform LULC mapping in a complex, subtropical environment. For this purpose, a specific data set based on SPOT6 satellite data was created and made available for the scientific community as an LULC benchmark in a tropical, complex environment using five representative areas of New Caledonia labelled by a human operator: four used as training sets, and the fifth as a test set. Several architectures were trained and the resulting classification was compared with a state-of-the-art machine learning technique: XGboost. We also assessed the relevance of popular neo-channels derived from the raw observations in the context of deep learning. The deep learning approach showed comparable results to XGboost for LC detection and over-performed it on the LU detection task (61.45% vs. 51.56% of overall accuracy). Finally, adding LC classification output of the dedicated deep learning architecture to the raw channels input significantly improved the overall accuracy of the deep learning LU classification task (63.61% of overall accuracy). All the data used in this study are available on line for the remote sensing community and for assessing other LULC detection techniques.

QADI as a New Method and Alternative to Kappa for Accuracy Assessment of Remote Sensing-Based Image Classification

“…In addition, the QADI was also able to deduce the cause of the classification error to be incorrect labeling. Figure 8 presents the QADI graph for the error matrix in Table 7 [47]. The Kappa value for this matrix is 0.48, and the computed QADI value is 0.45905 and lies within the red area, which indicates a very low confidence.…”

“…We also used five error matrices from earlier work and research literature as input for the QADI to examine its efficiency. Therefore, we considered error matrices for the accuracy assessment of the LULC classification using different data-driven approaches, namely OBIA (Table 6: [46]), deep learning (Table 7: [47]), and the three machine learning algorithms random forest, support vector machine, and artificial neural network (Tables 8-10: [48]).…”

“…The authors would like to thank the authors of the publications Naboureh et al [46], Rousset et al [47], Leeuwen et al [48] and for their support and for permitting to use their respective error matrices for comparison.…”

“…Error matrix for accuracy assessment of LULC classification based on deep learning (Rousset et al[47]). …”

AMM-FuseNet: Attention based Multi Modal Image Fusion Network for Land Cover Mapping

“…Thanks to their complementary features, multi-modal remote sensing imagery provides much richer information compared to single modality especially for land cover/use applications. However, in the literature, most of the land cover mapping papers still use single modality data [19][20][21]. Along with the technical developments in computational imaging and deep/machine learning research, the usage of multi-modal for land cover mapping data [22,23], despite being in early stages and insufficient, start to appear in some works in the recent years.…”