| The paper based on the Remote Sensing, Geographical Information System, Global Position System (shortened form is 3S) and the SPOT image, integrated with the samples investigation research and the field investigation, estimated the acreage of cotton in Xinjiang. At first, we reviewed and analyzed the development and questions of the 3S technology in crop acreage estimation, and then we introduced the principle and advantage of geometric correction using the GPS data, mainly on unsupervised classification, the minimum distance, the maximum likelihood and the Mahalanobis distance of supervised classification, and analyzed the principle and precision of classification. We estimated the crop acreage of different samples relying on the interpretation of the SPOT image in the different regions, different time and the field samples investigation and crop acreage statistics. We chose the Zepu county in Xinjiang as a development example, calculated the cotton acreage based on samples investigation and classification, and through interpolation method. The main purpose of this paper is to utilize the 3S technology, simplifies the complexly estimation technology of the cotton acreage in Xinjiang, provides the convenient, practical, economical large cotton acreage estimation technology. The main contents and conclusions are as follows:1. According to the spectral characteristics of the cotton and phenological calender of the crop, we educed that the optimal time of cotton recognition is the middle and last ten days of six month and the middle and last ten days of nine month.2. To the Aksu and Kashi region in Xinjiang. According to spectral effect of the SPOT image and the spectral characteristics of the cotton in Xinjiang, layer stack made different false color composite images using the SPOT2, 3 and 4 band. After comparisons and practical application, we found that the false color composite image by the 3, 4, 2 band is better than the image by the 3, 2 and 1 band.3. Tassled Cap and Principal Component Analysis are the powerful enhancement method dealt with the SPOT image. After the Principal Component Analysis, the image can clearly distinguish the distribution of cotton , compared with the field investigation data and the result of classification, the results supported the processing results, thus improved the precision and reliability of the classification of the crop acreage. Comparison the image by the PCA, the classification research of the cotton in Xinjiang is feasible and reliable. After tassled cap, the information ofthe image is more clear than the original image, but the image is not as good as the PCA.4. In this research, we attempt to dealt with the SPOT image' s geometric correction using the measured data of the GPS. The result indicated that the corrected image is not only high precision, but also the corrected method is convenient, which can have match the samples and the image quickly and accurately, and achieved satisfactory results, which increased the accuracy of cotton acreage estimation.5. In this paper, the images were interpreted by the four remote sensing classification method .By the evaluation of the classification accuracy and the analyse of statistical comparison, The minimum distance method is the best method to classify the cotton image, the result is in terms of the cotton distribution of the Xinjiang.6. Traditionally, the cotton area estimation in Xinjiang is situation with heavy workload and long duration, which will easily lead to the errors of the data and chaotic management, so we need systematized and standardized the flow of the cotton area estimation. In this paper, we systemized technical procedures, which can highly enhanced the precision and efficiency of cotton acreage estimation.
|
PDF Full Text Request