Liu Yongxue

Review of Remotely Sensed Imagery Classification Patterns Based on Object-oriented Image Analysis

With the wide use of high-resolution remotely sensed imagery, the object-oriented remotely sensed information classification pattern has been intensively studied. Starting with the definition of object-oriented remotely sensed information classification pattern and a literature review of related research progress, this paper sums up 4 developing phases of object-oriented classification pattern during the past 20 years. Then, we discuss the three aspects of methodology in detail, namely remotely sensed imagery segmentation, feature analysis and feature selection, and classification rule generation, through comparing them with remotely sensed information classification method based on per-pixel. At last, this paper presents several points that need to be paid attention to in the future studies on object-oriented RS information classification pattern: 1) developing robust and highly effective image segmentation algorithm for multi-spectral RS imagery; 2) improving the feature-set including edge, spatial-adjacent and temporal characteristics; 3) discussing the classification rule generation classifier based on the decision tree; 4) presenting evaluation methods for classification result by object-oriented classification pattern.

Extract residential areas automatically by New Built-up Index

Land covers in urban areas trend to change more drastically over a short period of time than elsewhere because of incessant urbanization. These changes are ideally monitored and detected from remotely sensed images as they are relatively up-to-date and give a panoramic view. In this paper we propose a new method based on New Built-up Index (NBI) to automate the process of mapping residential areas. The unique spectral response of land covers is most distinctive from one another for each cover in band 3, 4, 5. It takes advantage of the unique spectral response of residential areas and other land covers. This method has been successfully applied to extract the residential areas in Changzhou City, Jiangsu Province of China as a representative of plain areas by manipulating the spectral bands of TM imagery in 2007. The mapping results at a high accuracy. It acquires a high accuracy than NDBI. So it can serve as a worthwhile alternative for quickly extracting.

A Semi-automation Road Extraction Approach Based on Fast Marching Method and Mean Shift Algorithm

A method for extracting road from high spatial resolution remote sensing imagery (HSRRSI) based on Fast Marching Method and Mean Shift Method is proposed in this paper. Firstly, the user inputs a set of nodes of the road. Secondly, preliminary partition on the imagery is achieved by Mean Shift Method. Last, the roads are extracted between the nodes user has set by the implementation of Fast Marching Method The test results for HSRRSI of IKONOS demonstrates that the novel method is robust and is capable of coping with the partial occlusion and false connectivity.

An Intelligent Method of Detecting Multi-factors Neighborhood Relation Based On Constrained Delaunay Triangulation

Spatial neighborhood relation detecting is the basis of organization, query, analysis and reasoning of spatial data. For the spatial neighborhood relations of the geographic entities are contained in the Delaunay triangulation, in this paper, the spatial neighborhood relations between multi-factors (including points, lines and polygons) are intelligently detected based on the Constrained Delaunay Triangulation (CDT). This approach consists of steps listed below: A matched candidate points index is set up in order to reorganize the related data of multi-factors. Then the CDT, which is constrained by edges of lines and polygons, is established with the source of coordinates in point index. After coding the CDT according to a rule this paper proposing, neighborhood relation of geographic entities can be searched automatically. And a global neighborhood relation (including separation and neighbor relation) among multi-factors is automatically established by spatial reasoning. This intelligent method of spatial neighborhood relation detecting, which is no need for manual intervention and not limited between two kinds of geographic entities, has high precision and great feasibility in practice.