Study on Registration Technologies of Different Species Remote Sensing Images
|School||National University of Defense Science and Technology|
|Course||Information and Communication Engineering|
|Keywords||Heterogeneous remote sensing image Imaging Characteristics Image registration Roof Extraction FAST LSR Lidar ISEF|
Image matching is an important prerequisite for remote sensing image information fusion . With the development of remote sensing technology , the amount of remote sensing data and remote sensing sensor resolution has been greatly improved , more and more types of remote sensing imaging sensors . Comprehensive utilization of a target different bands of the electromagnetic wave radiation characteristics and a variety of sensor information to achieve a comprehensive understanding of the target characteristics , the need to ensure that the remote sensing image geometric registration . This paper first introduces the concept of image registration principle , the general classification of the image registration techniques and geometric transformation model , and then summarize the imaging characteristics of the heterogeneous remote sensing image registration techniques and a variety of sensors . This article describes the basic principles and common image pre-processing of remote sensing image imaging technology , a heterogeneous remote sensing image registration algorithm based on structural characteristics . Structural features mainly refers to the characteristics of the corner feature extraction using FAST algorithm as well as a straight-line support area (LSR) . For complex imaging of remote sensing images , the radiation characteristics and geometric characteristics change sensor types , the registration algorithm versatility , two registration programs are given according to the different properties of the image and the registration requirements of this article . The experiments show that the given algorithm has better registration effect . Lidar data with traditional remote sensing image registration problem , based the specific feature oriented Lidar heterogeneous image registration algorithm framework , and a Lidar depth image based on the infinite symmetric exponential filter (ISEF) roof extraction algorithm, and having a high positioning accuracy of the extracted algorithm proved by experiments . On the basis of the above algorithm , a heterogeneous remote sensing image barebones . Last reviewed the work done on this article and research , and the direction of future work prospects .