Land Use/Cover Change Detection Using Feature Database Based on Vector-Image Data Conflation

Hong Zhang + and Ning Shu
1 School of Remote Sensing and Information Engineering, Wuhan University,
129 Luoyu Road, Wuhan 430079, China

Abstract. Change detection in remotely sensed imagery is defined as the procedure of quantitatively analyzing and identifying changes occurred on the earth’s surface from remotely sensed imageries acquired at different times. Land use change survey with remote-sensed imagery has been one of the important methods for the land manage apartment to understand and accommodate land resources, and has attracted universal attention. As a key element for many applications of RS such as resource inventory, environment monitoring, update of fundamental geographical database, etc., change detection technique is of urgent demands and has great potential in scientific applications.         Conflation  is  the  process of combining the information from two (or more) geodata sets to make a master data set that is superior to either source data set in either spatial or attribute aspect. The objectives of conflation include increasing spatial accuracy and consistency, and updating or adding new spatial features into data sets. Based on the analysis and summarizations of researched home and aboard, the dissertation focused on Land Use/Cover Change detection using feature database of basic types based on vector-image data conflation, that is : Combining of  Land use map and RS image,  feature is extracted. This methodology belongs to “Feature class” of LUCC. It should be pointed out that the researches must be focused on the land use span other then traditional methods of the pixels. The main contributions of the study were summarized as follows: 1、Feature extraction based on land use span. The land use span is expressed by vector polygon along with raster region. First the spectrum feature database with histogram, texture and shape feathers of the span is formed. 2、Foundation and update of feature database. In detail, firstly, by means of the sample span according to land use map in time T1, the features of each type of the land use classes are obtained in time T1. Secondly, each sample are analyzed, if the index of regional similarity between the image span of T1 and T2 is accepted, the samples in time T2 could be remained, otherwise  the new samples around that sample are selected and are judged by the similarity between the samples of T1. 3、Change detection based on and feature database. Each span of T2 will be classified according to the minimum Euclidean distance to the T2 sample span accepted, and the corresponding land use type will be assigned to the current span. 4、Change information are extraction automatically based on Boolean operations. After classifications have been performed, the changed span were vectored, then the change information can be statistic through the different Boolean operations in GIS, and various change analysis can be made (i.e. urbanization and loss of the stew) The method is tested on the Quick Bird images of a district in Wuhan and the accuracy of the results is high as 85.7% (in loss of the stew) and 92.6% (in urbanization), and overall accuracy is 88.3%.

Keywords: land use map, sample span, vector-image data conflation, feature database, LUCC, accuracy analysis 

In: Wan, Y. et al. (eds) Proceeding of the 8th international symposium on spatial accuracy assessment in natural resources and environmental sciences, World Academic Union (Press).

HongZhang2008accuracy.pdf335.02 KB