Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/6169
Title: Investigation and remediation of False Topographic Perception Phenomena observed on lunar orbital imagery
Authors: Gao, Yang
Keywords: Remote sensing.
Hong Kong Polytechnic University -- Dissertations
Issue Date: 2013
Publisher: The Hong Kong Polytechnic University
Abstract: False Topographic Perception Phenomenon (FTPP) is a relief inversion phenomenon in remote sensing images and causes false perception problems, for example, craters appear as hillocks, mountains appear as valleys, and vice versa. Such images of the Moon suffer from serious FTPP problems. This is mainly due to the highly rugged and dust grayout surface, the absence of atmosphere eliminates the scattering effects, and the lack of familiar landmarks on the Moon, to supply visual cues. Correctly observing and understanding the terrain features on the Moon from lunar surface images is important for lunar exploration missions and various lunar scientific investigations. Therefore, this study systematically investigates the FTPP problem observed on lunar orbiter imagery (mainly based on the Chinese Chang'E-1 imagery) and develops methods to alleviate the FTPP problem on lunar orbiter imagery. Firstly, a systematic investigation of the FTPP problem observed on the Chang'E-1 imagery was performed. The Chang’E-1 images covering the whole lunar surface was examined, and terrain features (mainly craters) suffer from obvious FTPP problems were identified. Results revealed that the FTPP problem is positively correlated to the latitude of the terrain features, and majority of the objects suffering from the FTPP problem on the Chang'E-1 imagery are in the north hemisphere. For craters within a similar latitude range, the FTPP level is positively correlated with the depth-diameter ratio of the crater. Other factors related to the FTPP problem was analyzed, such as the sun elevation and azimuth angles, observation angles, and influence of shadows. Secondly, several traditional methods were investigated to alleviate the FTPP problem, which included the method of image rotation and the Digital Number (DN) invert method. Experimental results showed that these traditional methods have advantages as well as disadvantages when dealing with lunar orbiter imagery. Thirdly, this study develops two approaches for FTPP remediation on Chang’E-1 imagery. One is a wavelet-transform based approach. Digital terrain models of several craters were generated from the Chang'E-1 imagery based on Photogrammetric techniques, and SRMs with specific light directions were derived. Data fusion was then performed among the original image and the Shadowed Relief Model (SRM) using the Shift Invariance Discrete Wavelet Transform (SIDWT) method. After that, images free of FTPP can be obtained. The other is a rotation-invariant approach, in which an illumination balanced image was firstly generated and then a bright-darker depth map was incorporated to generate images free of FTPP. Experimental analysis using three different types of craters revealed that both methods are able to effectively correct the FTPP problem, while the wavelet-transform based approach is only recommended for circumstances with no image rotation. The rotation-invariant approach, however, is not subject to image rotation and is suitable for applications such as those using the lunar surface images for navigation or exploring lunar surface using software systems such as Google Moon.
Description: 111 leaves : ill. (some col.) ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577M LSGI 2013 Gao
URI: http://hdl.handle.net/10397/6169
Rights: All rights reserved.
Appears in Collections:Thesis

Files in This Item:
File Description SizeFormat 
b26160377_link.htmFor PolyU Users203 BHTMLView/Open
b26160377_ir.pdfFor All Users (Non-printable)3.92 MBAdobe PDFView/Open
Show full item record

Page view(s)

357
Last Week
3
Last month
Checked on Apr 23, 2017

Download(s)

147
Checked on Apr 23, 2017

Google ScholarTM

Check



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.