An Improved Method for Impervious Surface Mapping Incorporating Lidar Data and High-Resolution Imagery at Different Acquisition Times

Abstract

Impervious surface mapping with high-resolution remote sensing imagery has attracted increasing interest as it can provide detailed information for urban structure and distribution. Previous studies have suggested that the combination of LiDAR data and high-resolution imagery for impervious surface mapping performs better than using only high-resolution imagery. However, due to the high cost of the acquisition of LiDAR data, it is difficult to obtain the multisensor remote sensing data acquired at the same acquisition time for impervious surface mapping. Consequently, real landscape changes between multi-sensor remote sensing data at different acquisition times would lead to the error of misclassification in impervious surface mapping. This issue has mostly been neglected in previous works. Furthermore, the observation differences generated from multi-sensor data, including the problems of misregistration, missing data in LiDAR data, and shadow in high-resolution images would also challenge the final mapping result in the fusion of LiDAR data and high-resolution images. In order to conquer these problems, we propose an improved impervious surface mapping method incorporating both LiDAR data and highresolution imagery at different acquisition times in consideration of real landscape changes and observation differences. In the proposed method, a multi-sensor change detection by supervised multivariate alteration detection is employed to obtain changed areas and misregistration areas. The no-data areas in the LiDAR data and the shadow areas in the high-resolution imagery are extracted by independent classification yielded by its corresponding single sensor data. Finally, an objectbased post-classification fusion is proposed to take advantage of independent classification results with single-sensor data and the joint classification result with stacked multi-sensor data. Experiments covering the study site in Buffalo, NY, USA demonstrate that our method can accurately detect landscape changes and obviously improve the performance of impervious surface mapping.

Individual Tree Parameters Estimation for Chinese Fir (Cunninghamia lanceolate (Lamb.) Hook) Plantations of South China Using UAV Oblique Photography: Possibilities and Challenges

“…Another reason for the error is that the spatial information below a certain threshold height is masked during the algorithm execution, which causes the partial crown delineation to be small. Thus, scholars believe that spectral information provided by high-resolution images is more reliable than spatial information that can be masked [61]. In addition, the registration of images and LiDAR data will produce geographic errors.…”

Integração de Parâmetros Ambientais em Imagens Multifontes

“…A combinação de dados LiDAR em imagens multiespectrais foram identificados em estudos que se dedicam a classificação de uso e ocupação do solo, para melhorar o desempenho e compensar as deficiências um do outro (Rapinel et al, 2015;Morsy et al, 2017;Luo et al, 2018;Wu et al, 2018;.…”

Spectral-Spatial Classification Integrating Band Selection for Hyperspectral Imagery With Severe Noise Bands

“…Objectoriented and deep learning approaches have often been applied. For the object-oriented classification, it takes objects as the processing unit [22]- [25] to consider the spatial information. The objects are first generated by segmentation, such as the fractal net evolution approach (FNEA) [26].…”