Application Research On Forest Change Detection Using Multi-source Remote Sensing Data--A Case Study In Zhangpu County

Forest is an important part of the ecosystem in Fujian province. It is important to master the status and change of forest resource timely for reasonable exploitation of forest and its renewal.In this paper, integrating with GIS, GPS technology and groud survey data, multi-source remote sensing data, including multispectral TM/ETM+ data, high spatial resolution SPOT 5 PAN data and SAR data were used to monitor the forest change in Zhangpu County. The aim is to find a way to explore both the spectral advantage of multispetral data and the ability of high spatial resolution image toreflect details and update big scale map. SAR data may be utilized as supplement of optical data under bad weather conditions in austral areas.The main content of this paper is as follows.1. After analysis of TM/ETM+ data's advantage over the forest change detection, Tasseled Cap transformation, Principal Component Analysis and Normalized Difference Vegetation Index were chosen to enhance the vegetation spectral information. Expert classification system was adopted to extract the main tree species in the littoral shelter forest. The result showed that the expert knowledge is the key factor affecting the classification accuracy. Satisfactory results will be get by adjusting the variables and their threshold value of expert knowledge base. Study on the expert classification technology will promote the change detection technology with multiresoure.2. Based on the analysis of the main forest change types, a processing flow using multiresouce remote sensing data to detect forest changewas put forword, supported by GIS technology. The classification system and methods for forest resourcewere also described. The overall and main species changes of forest were achieved by hierarchical classification and thematic extraction. The result indicated that the forest resource in Zhangpu County had changed a great during 1997-2003.These changes include both positive changes from nonforest to forest and negative changes from forest to nonforest. The top ten changes by area were as follows: plantation to cut area, forest toopen forest, forestto cut area,non forest to forest, barren land to forest,plantation to plowland,forest tobarren land,plantation to barren land,nonforest landto forest and open forest to forest. Based on the forest changes ,some suggestion on the management and ecoconstruction were put forward.3. Four experiments on the application of SAR data to extract forest resource information were carried out:(l)use SAR data only;(2)PCA transformation was carried out to TM data, then , the top three PCs were combined with SAR magnitude image and SAR texture image;(3) All TM bands (TM6 excluded) and SAR data were combined together, and then PC transform were adopted;(4) SAR data(2000) were merged with TM. The result showed that experiment (1) performed bad, and experiments 4 didn't get the expected good result. Only experiments (2) performed better in the discrimination between different types of forest.