The Research On Monitoring Of Conversion Of Cropland To Forest With Remote Sensing Classification

Since 1999, the implement of the project of conversion of cropland to forest has achieved good ecological, social and economical effect. However, so many of project quantities that they bring problems about quality acceptance and the monitoring of implementary effect of the project. The technology of Remote sensing and Geographical information system becomes effective means for the monitoring of conversion of cropland to forest.The paper takes Kangping county , Liaoning province for case to research for three years from 2003 to 2005. Remote sensing images used in this study are the SPOT-5 images which were acquired in 2003 and in 2005,respectively . We could achieve available images which are used to extract changed informat- ion by approaches with image preprocessing, image geometric rectification, ortho- rectification, image fusion, image enhancement, image mosaic and image clipping. Changing information auto-detection technique is the most critical and complex aspect in the forest resources dynamic monitoring. First, both SPOT-5 images are auto-classified by computer with the method of supervised classification and then to assess classifica- tion accuracy. The result are 94.4% and 92.6%, respectively. Second, forest resources changes are extracted by post-classification, then assessing the accuracy of the classification results, the result is 87.4%, up to requirements. At last, accordi- ng to the classification results, we could generate the area transfer matrix and estimate the area quantities of forest resources changes in Kangping county. As to the monitoring of the project of conversion of cropland to forest, at first, we stack the layers of designed project subcompartments, then base on the image of forest resources changes to determine the sources of changes. With transfer types of land sort, we select subcompartments samples and check to fields. The sampling inspection shows that the correct rate of total subcompartment is 94.12%. Finally, we statistic area quantities of each transfer types and assess the implementary effects of the project of conversion of cropland to forest.