Pixel based and texture analysis to integrate remote sensing for efficient describe structures in urban land uses

Resumen

Today, nearly half of the world's population lives in cities. In developing countries, people are deserting rural areas while population is rising rapidly. In less than 20 years from now, these two factors will combine to drive over two billion people into urban areas, which in some cases are already overcrowded. Most urban growth falls outside formal planning controls, thus increasing economic and social pressures and exacerbating health and hygiene problems. The main advantages of Remote Sensing with GIS are to bring parties together, to spread and improve idea, to support the decision-making process, and to evaluate projects. Therefore an automatic or semi- automatic land-use mapping approach would be preferred to support feature extraction and land-cover and land-use classification, a Pixel-Based image approach developed and investigated in this research, where image objects are defined based on the size, shape and pattern. Topological relations between image objects at different abstraction levels are defined and extracted based on image regions. In turn, structural analysis and spatial clustering or spatial units of land use can be extracted based on the texture analysis approach, which is essential to accomplishing land-use classification. The proposed concepts and approaches will be tested on two case study areas. The first test site is the Metropolitan Region of Barcelona. The second case considers a complex mixed between the land use and cover areas which find in the Metropolitan Region of Madrid, the data for these densely built-up urban areas are SPOT images. These two different cases were also selected with a view to including different land-use types and spatial patterns in the investigation and examining the effectiveness of different data combinations. Digital images, particularly those from remote sensing technology, have become an important source of spatial information. This research note demonstrates the methodologies and techniques of extracting thematic information from digital images. The discussions are focused on the techniques of image enhancement, interpretation and auto-classification using black-and-white (or single band) images and multi-spectral images. Methods and techniques used for integrating digital images with spatial information systems are also discussed. The confuse class problem becomes a real issue when we use high-resolution data (2.5m to 10m) for urban areas, although it can be ignored when using coarse resolution data or dealing with non-urban sites. This problem is difficult to solve using per-pixel approaches alone. This problem has been largely solved by using spectral data and texture analysis in our methodology. Although the proposed modifications improve the land-cover classification accuracy of the Pixel Based, we consider the attainable results insufficient for a detailed urban land-use classification because of the complexity of urban environments. Finally, Remote sensing technology was used to extract urban-used land information and the urban pixels in the classification image were directly used to calculate the urban compactness in both 2D and 3D directions, which wil l be more proper. Keywords Remote sensing, GIS, Texture analysis, Feature Extraction, LiDAR, Compactness, change detection, urban agglomerations and delimitation.