2019_Segmentation Technique for Flood Detection System Using UAV Images
| Format: | General Document |
|---|
| _version_ | 1860798037423554560 |
|---|---|
| building | INTELEK Repository |
| collection | Online Access |
| collectionurl | https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 |
| copyright | Copyright©PWB2025 |
| country | Malaysia |
| date | 2019-09-29 |
| format | General Document |
| id | 15586 |
| institution | UniSZA |
| originalfilename | SEGMENTATION TECHNIQUE FOR FLOOD DETECTION SYSTEM USING UAV IMAGES (MASTER_2019).pdf |
| person | Nor Shuhada binti Ibrahim |
| recordtype | oai_dc |
| resourceurl | https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15586 |
| spelling | 15586 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15586 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu Faculty of Innovative Design & Technology English application/pdf 1.5 Universiti Sultan Zainal Abidin UniSZA Private Access Universiti Sultan Zainal Abidin SAMBox 2.3.4; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) Copyright©PWB2025 187 2019-09-29 SEGMENTATION TECHNIQUE FOR FLOOD DETECTION SYSTEM USING UAV IMAGES (MASTER_2019).pdf Segmentation Techniques Floods Detection Flood control Image processing Natural disasters 2019_Segmentation Technique for Flood Detection System Using UAV Images Flood is a disaster that regularly occurs almost yearly. In Malaysia, Department of Irrigation and Drainage (DID) monitored this phenomenon based on few techniques; include the visual surveillances which are from closed-circuit television (CCTV) web camera and Synthetic Aperture Radars satellites (SAR). Yet, CCTV web camera is only covers in specific areas and requires more installation in residential areas. This continuous monitoring sometimes provides unclear information. Meanwhile, for satellites imagery DID admit with high operation cost and time route not at the peak flooding, it is difficult for the agency to obtain real-time information. Both techniques are based on qualitative evaluation, which is required others supporting data to verify, need special officer from DID and sometimes need to verify on flood scene. Thus, the aims of this study are to study the performance of color model and segmentation technique that can be used to detect the flood in Unmanned Aerial Vehicle (UAV) images. Secondly are to develop flood detection algorithm by proposing the Percentage Flood Discovery (PDF) formula. Then, by evaluate all the component of color model, segmentation techniques and PFD to ensure that this proposed study able to detect and clarify the UAV images as flooded or non-flooded in details based on qualitative and quantitative evaluation. To overcome these issues, the study conducted on two different situations; first, getting images from flood simulation model and second, attaining river images using UAV as flood detection. Firstly, by investigated grey images with two (2) color models namely RGB and HSI to be fed into the image segmentation algorithms conducted by three (3) image segmentation techniques namely Otsu‟s Thresholding, K-mean Clustering and Region Growing. To evaluate the segmentation performance, four (4) performance metrics which are Accuracy, Precisions, Recall and F1-Score applied. For pilot test, proposed flood simulation model (FSM) was applied and the Percentage Flood Discovery (PFD) created based on the overlapping images with dimension of 50x50 pixels that also applied in UAV images. Therefore, flooded areas with dimension less than 50% were not included in the flood classes. All the methods and data analysis were carried out in MATLAB simulator. For FSM, the average highest accuracy for color models obtained from green color of RGB is at 93% from region growing segmentation. Whilst, for UAV image cases, the performance was achieved also by region growing segmentation with accuracy at 88% from red color model; K-mean clustering and Otsu‟s Thresholding are at 76% obtained also from red color model, RGB and 77% from blue color model. The overall proposed image segmentations were able to detect the flooded areas. The selection of color model being the most important step in this segmentation. Since the land area of FSM was created in green color, there are a need to prove using river images to demonstrate the ability of UAV imagery and image segmentation process as a platform for flood detection and accurately extraction of inundated areas. Thus, with consistent flying altitude, it certainly helps DID to recognize flood areas during rainy season and forecast future mapping process. Nor Shuhada binti Ibrahim Dissertations, Academic Thesis |
| spellingShingle | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| state | Terengganu |
| subject | Flood control Image processing Natural disasters Dissertations, Academic |
| summary | Flood is a disaster that regularly occurs almost yearly. In Malaysia, Department of Irrigation and Drainage (DID) monitored this phenomenon based on few techniques; include the visual surveillances which are from closed-circuit television (CCTV) web camera and Synthetic Aperture Radars satellites (SAR). Yet, CCTV web camera is only covers in specific areas and requires more installation in residential areas. This continuous monitoring sometimes provides unclear information. Meanwhile, for satellites imagery DID admit with high operation cost and time route not at the peak flooding, it is difficult for the agency to obtain real-time information. Both techniques are based on qualitative evaluation, which is required others supporting data to verify, need special officer from DID and sometimes need to verify on flood scene. Thus, the aims of this study are to study the performance of color model and segmentation technique that can be used to detect the flood in Unmanned Aerial Vehicle (UAV) images. Secondly are to develop flood detection algorithm by proposing the Percentage Flood Discovery (PDF) formula. Then, by evaluate all the component of color model, segmentation techniques and PFD to ensure that this proposed study able to detect and clarify the UAV images as flooded or non-flooded in details based on qualitative and quantitative evaluation. To overcome these issues, the study conducted on two different situations; first, getting images from flood simulation model and second, attaining river images using UAV as flood detection. Firstly, by investigated grey images with two (2) color models namely RGB and HSI to be fed into the image segmentation algorithms conducted by three (3) image segmentation techniques namely Otsu‟s Thresholding, K-mean Clustering and Region Growing. To evaluate the segmentation performance, four (4) performance metrics which are Accuracy, Precisions, Recall and F1-Score applied. For pilot test, proposed flood simulation model (FSM) was applied and the Percentage Flood Discovery (PFD) created based on the overlapping images with dimension of 50x50 pixels that also applied in UAV images. Therefore, flooded areas with dimension less than 50% were not included in the flood classes. All the methods and data analysis were carried out in MATLAB simulator. For FSM, the average highest accuracy for color models obtained from green color of RGB is at 93% from region growing segmentation. Whilst, for UAV image cases, the performance was achieved also by region growing segmentation with accuracy at 88% from red color model; K-mean clustering and Otsu‟s Thresholding are at 76% obtained also from red color model, RGB and 77% from blue color model. The overall proposed image segmentations were able to detect the flooded areas. The selection of color model being the most important step in this segmentation. Since the land area of FSM was created in green color, there are a need to prove using river images to demonstrate the ability of UAV imagery and image segmentation process as a platform for flood detection and accurately extraction of inundated areas. Thus, with consistent flying altitude, it certainly helps DID to recognize flood areas during rainy season and forecast future mapping process. |
| title | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| title_full | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| title_fullStr | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| title_full_unstemmed | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| title_short | 2019_Segmentation Technique for Flood Detection System Using UAV Images |
| title_sort | 2019_segmentation technique for flood detection system using uav images |