Abstract:
During the last few years, site investigations have been an important aspect to the civil construction such as dams, towers, bridges …etc. It has been a challenge to find the most appropriate technique for studying soil and rock properties effectively and at the same time reducing the amount of field work such as number of test pits or boreholes, sampling in addition to laboratory analysis, which in turn lead to decreasing the cost of the specific project and achieve or performed in short time. The objective of the present study is to test the viability and the usefulness of Remote sensing, GIS technology and limited ground survey for soil classification
The study area is located in northern Sudan about 200 km downstream of Dongola, immediately upstream of Lake Nubian, and about 700 Km from Khartoum. The area is characterized by one of the most extreme desert climates in the world. The main elements of the geology include high-grade gneisses structurally overlain by a sequence of green schist metasedimentary /metavolcanic arc assemblage. These rocks are intruded by I-type and A-type granitites, with some tectonic mélanges representing dismembered ophiolitic fragments. Cretaceous Sandstone disconformabley overlies the previous rock units.
Different image processing techniques were applied in this study. Some of which are directed towards the image preparation. Others are used to enhance the visual interpretability of the images. Landsat ETM+ image was digitally enhanced and interpreted; integrated with the results of soil laboratory analysis of 18 samples in the GIS, and soil types were delineated to facilitate the production of the final soil map of the study area.
Ratio digital image technique pointed out to the existence of various soil types with clear boundary appeared in totally different colours. Transformation techniques such as PCA and RGB to HIS were found to be of great value when conducting soil classification. The resulting images from these transformations were found to be the best among the enhance images and produced reasonable results when applying image classification upon them.
Automated image classification was not completely produced the soil map, rather it was obtained by integration various enhanced Landsat images and laboratory analysis. However, care must be taken when conducting soil classification using satellite image, since certain soil types have the same reflectance properties in some band combinations, while others have similar signatures to some rock types.
