Synergetic Effect of Complementary Nature of Hyperspectral and LiDAR Data for High Performance LULC Classification

Smart cities are being developed by using well-planned schemes based on a wide variety of data from different sources. Remote sensing is a highly valued technology that provides multimodal information using Hyperspectral and LiDAR data. Such sensor responses are capable of sketching footprints for smart city planning. We have shown the synergetic effect of complementary nature of both Hyperspectral and LiDAR data. It facilitates the land-use/land-cover (LULC) classification to provide precise footprints that subsequently aids on smart city planning. We have demonstrated the proof of concept using datasets provided by National Ecological Observatory Network (NEON). The geographical location covered in the scene being captured in the form of Hyperspectral Image consisting three major classes, viz., vegetation, soil, and road. The results of experiment have shown an overall classification accuracy of 98.61% with synergetic effect along with the performance improvement of 1.96% and 5.39% with respect to Hyperspectral and LiDAR data, respectively. In this experiment, a widely popular neural network architecture Convolutional Neural Network (CNN) has been used as the classifier for performance assessment.