Extraction of Four Types of Urban Ground Objects Based on a Newly Created WorldView-2 Multi-colour Spectral Index

Abstract

Fast and accurate extraction of urban ground objects can ensure timely urban environmental monitoring. However, the current technology used to extract urban ground objects has many steps, slow speed and low practicality. In view of the shortcomings of current research, this study takes WorldView-2 as an example, collects typical image categories of urban imagery, conducts spectral curve fitting and characteristic analysis, uses a combination of the addition, subtraction and multiplication operations in some key bands to enhance the information of the target objects and constructs some spectral indices that are beneficial to the extraction of urban red, white, blue and cyan ground objects using the threshold method. The newly created spectral indices can effectively extract four colours of urban ground objects. Among them, the red and blue ground objects have good extraction results, whereas the white and cyan ground objects have poor extraction results. Results show that the new spectral index threshold method can rapidly extract the target ground objects in urban images by colour, which has good practicability.