| Surveying Land-use is the basis for land management, and is the actual source for the accurate data that present land-use at all levels in our country. Traditional investigation means about land-use require a lot of manpower and material resources, but inefficient. Developing with the remote sensing, the geographic information system and the technology of global positioning system, and the emergence of high-resolution satellite images, furnishing a lot of efficient and reliable message source and advanced technologies for land resources survey.This paper focuses on using the technologies of RS (Remote Sensing), GIS (Geographic Information System) and GPS (Global Position System), according the high-resolution satellite images as a major source of data,to introduce the work process and technology issues in land-use surveys. The result of the study will provide practical techniques and methods for land-use surveys based on high-resolution satellite images, land-use changing and arranging the land-use map of 1: 10,000.In this study we adopt SPOT-5 satellite images with 2.5 m spatial resolution in 2005, carrying into land-use survey execution in local regions of Jilin City for instance. First enhancing the image and correcting the geometric in ERDAS software environment, mainly adopt the image enhancement techniques to enhance the imaging characteristics of different types of terra; then gather the reference points with static difference of GPS, used polynomial correction method to geometric accurate correct the images that had been enhanced; and then produced the striograph of 1 : 10,000. Separate print output for the use of external investigation. Through interpretation signs established base on fieldwork, according to the latest promulgated the "National Land Classification (Trial)" to establish the research section's land use classification system, supervised classification with ERDAS image processing system, to get the delineation of land utilization and its attributive classification information, contrasting the software classifications with external investigation to evaluate the assoeted precision. Since that using the SUPERMAP software to layered vector quantize other geographical factors, together with the establishment of its attribute table: and analysis, storage and preparation of land-use maps.Evaluation the assoeted results as follows:Using the maximum likelihood method supervised and classified the images with the ERDAS software on the computer. Due to mainly country settlement place, infield and its subsidiary facilities in the region, the researching area is divided into 10 main types. Such as Paddy field 1, Paddy field 2, Drains, The residents, Dry, Roads, Rivers, Railways, Pit, The independent mines. The reason of the paddy field divided into two kinds is that the two types of paddy fields have a greater difference after the harvest in action spectrum information.After classifying the software and correcting by eyes, the assoeted accuracy overall the region is 88.28%, seeing from the nicety of grading, Paddy field 2, Pit, and Paddy field 1 have the highest precision, respectively 100%, 100%and 98.31% ; Drains, The residents, Dry, Roads and Railways have higher precision, respectively 89.47%, 88.46%, 87.23%, 88.89% and 80.95%, The independent mines, rivers have the lowest accuracy, 66.67%, 30.77% respectively.The main reason is SPOT-5 Imaging Paddy fields, Pits, Dry lands and The residents share a larger percentage of the area, have a evident characteristic, easy to distinguish. The glebe such as Infields, Drains, Roads, Pits, etc. usually planning as artificial, so different from the areas around, the type of The residents and Infields often demarcate with canals and roads, easy to distinguish.The independent mines are small in the area, particularly as the traditional rural mine enterprises such as a brick yards, dilapidated serious, immingled with the residents, the boundaries are not clear and hard to judge. In addition, the reason of the river's classification accuracy is the lowest is that the satellite images taken at the time of September, when the rivers on the high flow period with generous water, reading the images identified clearly, and the outdoor operation doing in May that the river's dry season, when the rivers on the low flow period with ungenerous water, impacted the accuracy of reading and investigating the data of outdoor operation, thereby affecting the classification accuracy.In addition, the enhanced disposal of amalgamated image exist errors, the result will be touched. As the images of SPOT-5 despite have a high spatial resolution, adapts to read out by visual, but not high spectral resolution. In some of the more complex areas, the type of resolution is not very high. Finally, as outdoor operation's judge and read errors also effect the classification. The above reasons also caused some lower accuracy the classification results.The findings of the study are summarized as follows:1.The traditional land use changing survey is the labor-intensive, the high cost and timeliness poor. It is difficult to meet the needs of socio-economic development and sustainable land use. The technology of Remote sensing have a good many virtues, such as macroscopical, dynamic, convenient, and can be duplicated and low cost, and its application to land-use surveys has become a general trend.2. SPOT-5 panchromatic and multi-spectral imaging not only retains the spectral information of original multi-spectral images but also improved spatial resolution images. After the amalgamation the spatial resolution image reached 2.5 m, greatly enhanced the definition, make the three-dimensional feeling strong, rich color levels, increase the cost performance. The precision entirely meet the need of 1:1 million land-use surveys, which will able to be a land-use survey map and heighten the fruit's quality in the outdoor operation investigation.3. When surveying land-use, we had better chose the images of the same phase. These images will increase the accuracy of land utilization and the interpretation precision in outdoor operation.4. The selective areas of this test have a simple landform, a sharp-cut style and a clear delineation. The text makes use the experience of operating personnel to participate in classification, and make full use of the spatial information of the rich images. These all can improve the precision. But if the selective areas are complicated terrain features and fuzzy character, the interpretation precision must be reduced. Therefore, when we use SPOT-5 Image to survey land-use, the investigating work of outdoor operations can not be replaced by the classify work of official operations, but can be assisted, or using it to monitor land-use changing.5. In some regions where conditions permit, using the image of recent time, will improve the accuracy of land-use surveys in outdoor operation, and then improve the rogatory precision.6. In the course of the investigation some kinds of terra have a big difference with its image, such as the glebe instead of the paddy field, the forest instead of the infield, the position of these areas can not be distilled exactly, and we must conduct field measurements in the field for operations.
|
PDF Full Text Request