Research on land Resource by Remote Sensing in inner Mongolia, China
Zhao Ji
Beijing Normal University Beijing 100875, People’s Republic of China
Introduction
Remote Sensing technique is a very effective method for resource investigation of large areas like The Inner Mongolia Autonomous Region. Basic Information of this area is provided by the MSS, TM, RBV, images of Landsat, the HRV image of SPOT and the image from domestic satellite. Some important specific information needed for this research, such as grass community, crop land, discarded crop land, herbage land, desertificated land, saline – alkalized land, various degradated land are provided by image processing techniques including the application of computer system and optical instruments. On the basis of the information checked by ground truth, field study and terrain analysis, a map to describe the present status of land use in this are with a scale of 1:1500000 is compiled.
Physical and Environmental Background
The Inner Mongolia Autonomous Region is located in northern China with an total area o 1,183,000 km2.
The Inner Mongolia Plateau having been raised a very long time ago and continuously eroded every since, is of an average elevation of 1000 to 1400 meters.
In winter it is cold and windy, in summer the area is affected by the monsoon originating from the Pacific Ocean to some extent. The monsoon is somewhat weekened by the time it reaches the plateau, so the annual precipitation is as low as 150 – 400mm. however, there are significant variations from year to year, depending on the force of the monsoon.
Under the influence of semi-arid climate, the predominant vegetation is steppe, with deserte-steppe gradually changing to desert in the western part of the plateau. The eastern part is the most fertile, especially in the DA Hinggan Mountains. The dominant vegetation is coniferous forest and mixed conifer/broad – leaved forest which stretches over the tops of the mountains.
A small part in the south is used as agricultural land to meet the need of food for the residents. Most of the area is used as pasture land constituting about 55.47% of the total area.
Methodology
1. Material
The data for this study were mainly derived from the CCT and FCC of MSS, TM imagery. However, other imagery and maps, such as the RBV, SPOT, Chinese scientific experiment satellite images and air photos, topographic maps, soil and geologic maps were also used as ancillary data sources.
2. Digital image processing
To improve the precision of the classification map and provide accurate data for the quantitative study of the project, some CCT of Landsat MSS, TM images were processed by IIS computer system using the
methods of supervised classification, exponential and level-slice. As a result, much information was extracted as the basis for this study.
The supervised classification was done in three steps. First, to enhance the information of vegetation and reduce the effect of shadow of landform, we made a ratio for all the four bands of MSS image. They are band 7/band 6, band7/band5, band7/band4, band6/band5, band6/band4 and band5/band4, which were called X1, X2, X3, X4, X5 and X6 respectively. Second, to reduce the amount of data used and make good use of the information extracted by ratio process, we used principal component analysis for all the six images. The result of this procedure are as follows :
So, we made a false color composite with PC11, MSS5 and MSS7. Finally, the supervised classification was conducted using the minimum distance option. Twenty two samples were selected as the training areas according to the field work and eleven types of land were formed automatically as the result of the method.
In FCC, sandland and saline-alkalized land have such similar photoelements. To separate them in image clearly, MSS4 was processed using exponential transformation option. After doing “Scale” obtain for the final image, the density value of sand land became, 85, saline-alkalized land 265 and other features 0 in this image.
One of the important characteristics of desertificated land is the decrease of vegetation coverage. The image of (MSS7 — MSS5) / (MSS7 + MSS5) is called the “Vegetation Index Image”, because the value of density is clearly related to vegetation.
The lower the cover of vegetation, the less the value of density in the image. So, by processing image of (MSS7 — MSS5) / (MSS7 + MSS5) with level – slide option, the information of different degrees of decertified land was obtained.
Calculation of the areas
To get the accurate values of the areas of various types of degraded land, the final classification map of land degradation type was sliced along the boundaries of each type and every of slices was weighted. The formula to get the value of the area of each type is as follows:
