Full text
| |
| Author(s): |
Guilherme Pina Cardim
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | Presidente Prudente. 2019-11-14. |
| Institution: | Universidade Estadual Paulista (Unesp). Faculdade de Ciências e Tecnologia. Presidente Prudente |
| Defense date: | 2019-09-17 |
| Advisor: | Erivaldo Antonio da Silva; Ignacio Bravo Muñoz; Mauricio Araujo Dias; Alfredo Gardel Vicente |
| Abstract | |
Digital image processing (DIP) consists of an area of great scientific interest in different areas. In Cartography, the DIP is widely used in remote sensing studies to extract cartographic features of interest present in orbital images. Among the cartographic features, the detection of road networks has become of great scientific interest, since it can provide accurate and updated information for urban planning, for example. In this sense, the scientific literature has several works proposing different methodologies of extraction of road networks in orbital images. It is possible to find proposed methodologies based on fuzzy logic, edge detector and growth by region, for example. However, the existing studies focus on the application of the extraction methodology to certain areas or situations and use orbital image cuts in their studies due to the large amount of information contained in these images. In addition, the technological advance has allowed the acquisition of remote sensing images with high spatial, spectral and temporal resolutions. This fact produces a large amount of data to be processed during studies developed in these images, which results in a high computational cost and, consequently, a high processing time. In an attempt to reduce the response time of the extraction methodologies, the developers dedicate efforts in reducing the complexity of the algorithms and in using some available hardware resources suggesting solutions that include software and hardware processing. Therefore, the present work proposes a methodology for the extraction of different types of road networks and orbital images of high spatial resolution, without the need to generate cutbacks. In this sense, the proposed methodology for road extraction is based on the algorithm of region growing, for which a new implementation was proposed in order to allow its execution in a GPU platform with the intention of obtaining better performances. The results of the extraction were statistically evaluated by metrics defined in the literature, presenting satisfactory results, being verified a greater difficulty to delimitate urban roads present in complex scenes. In terms of performance, the use of the proposed algorithm of region growing allows to perform the extraction methodology with 20% less time in the majority of the images tested. In addition, the extraction methodology was applied to the GPU platform to compose the proposed co-desing system. The behavior and performance of the extraction methodology was evaluated in two distinct GPUs, obtaining advantages in the execution of large images. However, when using the GPU platform to process smaller images, two different scenarios were obtained according to the evaluated GPU. For the most powerful GPU, the processing time was advantageous for the vast majority of the images, but when using the less powerful GPU, the cost of transporting the image between the computer and GPU memories did not allow the results to be obtained more quickly. Thus, this research contributes to the scientific literature by proposing a methodology for the extraction of different types of road networks, which can be applied to a complete remote sensing image of high spatial resolutions without the need to split it. In addition, a co-design system was generated and evaluated in relation to the advantages and disadvantages obtained in the application of the proposed extraction methodology in two different GPU platforms. (AU) | |
| FAPESP's process: | 14/24392-8 - Proposal of co-design platform for processing remote sensing images |
| Grantee: | Guilherme Pina Cardim |
| Support Opportunities: | Scholarships in Brazil - Doctorate |