Land cover and land use characterization with geobia in the Pitangui River Basin area, Parana-Brazil
GEOBIA - Geographic Object-Based Image Analysis - is considered to be a technique that takes advantage of several Remote Sensing dimensions – like spectral, spacial, morphologic and contextual information – together with functionalities of Geographic Information Systems (GIS) as well. GEOBIA unit of...
Lưu vào:
| Tác giả chính: | |
|---|---|
| Định dạng: | BB |
| Ngôn ngữ: | eng |
| Thông tin xuất bản: |
2020
|
| Chủ đề: | |
| Truy cập trực tuyến: | http://tailieuso.tlu.edu.vn/handle/DHTL/4788 |
| Từ khóa: |
Thêm từ khóa bạn đọc
Không có từ khóa, Hãy là người đầu tiên gắn từ khóa cho biểu ghi này!
|
| Tóm tắt: | GEOBIA - Geographic Object-Based Image Analysis - is considered to be a technique that takes advantage of several Remote Sensing dimensions – like spectral, spacial, morphologic and contextual information – together with functionalities of Geographic Information Systems (GIS) as well. GEOBIA unit of analysis is an image object that obtains complete information. It includes texture, shape and spatial relations to neighboring objects and ancillary spatial data to different resolutions. In this study, it was sought to identify – in the spatial, spectral and texture descriptor group generated with GEOBIA – those that most represent the study unity. The technique was applied in part of the scene 221/77 of Landsat 5 TM satellite from year 2010 in color composition R5G4B3. This area belongs to the Pitangui river basin located at the center-east portion of the State of Paraná - Brazil, between coordinates 600.987m E and 7.240.620m N to 614.973m E and 7.231.568m N and zone 22S. Segmentation tests were done adjusting scale and merging levels according to the proposed objective (analysis of land use). Using a group of 35 descriptors with spatial, spectral and texture characteristics, Principal Component Analysis (PCA) and clustering were respectively applied with the aim of decreasing dimensionality of descriptors and grouping similar ones according to the shortest euclidean distance. From 14 resulting descriptors analyzed in the dendrogram generated in the clustering procedure, only compactness and solidity were grouped in the shortest euclidean distance. In the descriptors graph, maximum and minimum values for compactness and solidity have presented linear increase enabling the previous analysis. |
|---|