| Author: | Chin, Sze Wing |
| Title: | A comparison of change detection techniquies to detect illegal landfilling in rural areas : a case study in Hong Kong |
| Other Title: | Corrected title : A comparison of change detection techniques to detect illegal landfilling in rural areas : a case study in Hong Kong |
| Advisors: | Zhu, Xiaolin (LSGI) |
| Degree: | M.Sc. |
| Year: | 2019 |
| Subject: | Dust storms -- Remote sensing Aerosols -- Remote sensing Atmosphere -- Remote sensing Artificial satellites Hong Kong Polytechnic University -- Dissertations |
| Department: | Faculty of Construction and Environment Department of Land Surveying and Geo-Informatics |
| Pages: | viii, 40 pages : color illustrations |
| Language: | English |
| Abstract: | Illegal landfilling is a big headache for rural areas. As there is less and less available urban land, rural areas becomes target for development. In recent years, there have been growing incidents in which vegetation lands have been illegally filled up and/or with vegetation removed prior to obtaining government approval. When natural vegetation land is destroyed, it is not easy to restore because the soil has become degraded through loss of nutrients. Also, the speed of paving the land is fast and the location of paving is remote, making it extremely difficult to detect landfilling sites without complaints from the public. How to detect landfilling sites proactively is therefore important to curb this illegal action to destroy the rural and natural environment. Remote sensing techniques, ranging from the classic post-classification analysis and common image differencing such as subtraction of Normalized Difference Vegetation Index (NDVI) as well as Change Vector Analysis (CVA), have been shown effective for detecting changes in land-use/land-cover on a regular basis. The paper aims 1) to study the usability of the three common kinds of change detection techniques for landfilling identification, 2) to compare the performance by investigating accuracy for each of them, 3) the effect of fill materials (e.g. bare soil, concrete, asphalt, etc.) and land cover (i.e. homogenous and heterogeneous) on the change detection techniques. Change detection was employed in rural areas of homogenous land cover and heterogeneous land cover with landfilling sites of different fill materials in Hong Kong, using two sets of Sentinel-2 images acquired in the same month before and after a year. Techniques for comparison include 1) post-classification analysis by Maximum Likelihood Classification (MLC), 2) difference of NDVI and 3) Improved CVA. Post-processing procedure was then applied to filter undesirable pixels. Accuracy assessment was based on true positives (TP), false positives (FP) Kappa coefficient (κ) and User's Accuracy (UA). All three methods have the capability of detecting landfilling sites. The NDVI differencing approach gives the highest accuracy. The traditional post-classification analysis performs with the lowest accuracy. Improved CVA, which applied supervised classification and image differencing, shows moderate results. Due to the comparable accuracy of these three techniques, therefore, they can only be used for an initial screening aimed to detect suspicious areas. Moreover, results indicated that homogenous land cover can detect better than heterogeneous land cover. In addition, fill materials affects the results too. Landfilling sites making up of bare soil are easier to be detected than those of concrete and asphalt. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 991022385341203411.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 1.9 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/10557