|
Woodland with people's production and life are closely related, with clean air, water conservation, regulating such an important role in the ecological balance. Thus, in the preparation of urban planning, it is required to strengthen the protection and scientific development of forest resources, promote economic, social and environmental harmony. However, with the accelerated process of urbanization, forest resources have been destroyed, real-time monitoring forest change is essential. Remote sensing has the weather, timeliness, etc., remote sensing thematic information extraction has become the primary means of access to information. Currently most of the only green spaces (including woodland, grassland, etc.) were extracted, and the visual interpretation or use more traditional pixel-based approach, there is no separate extraction of forest information. This paper summarizes the research status green information extraction based on remote sensing based on three different resolution images (ETM, SPOT5, QuickBird images), object-oriented information extraction technology to achieve the forest information extraction. This article outlines the history and development of remote sensing principles, summarized information extraction technology and remote sensing information extraction Green Research. Introduced object-oriented information extraction method of the study area, as well as ETM, SPOT5, QuickBird image features, and images were orthorectified, registration, fusion and other treatment. An area of Kunming, for example, by analyzing the woodland in three typical characteristics of the image and the optimal segmentation scale, object-oriented information extraction method of woodland, and three kinds of precision extraction results were evaluated. Meanwhile, the three types of pixel-based images using traditional methods (maximum likelihood method) to extract information woodland and conducted accuracy assessment, comparative analysis of the traditional pixel-based methods (maximum likelihood method) extraction results and object-oriented methods The extraction results to arrive at useful conclusions, object-oriented approach in high resolution remote sensing image information extraction has obvious advantages. The woodland information extraction results are applied to urban planning, by calculating forest cover, combined with gradient information, designated a \urban planning decisions, to protect the interests of science, developing forest resources. In this paper, an area of Kunming, for example, based on high-resolution remote sensing images, the use of object-oriented approach, by selecting the optimal segmentation scale woodland and woodland typical of scientific analysis to achieve a woodland information is accurate, efficient and economical extraction, while information to achieve a quick update woodland.
|