Gwangsu HO (I), Akira KAWANAKA (I), Mikio TAKAGI (II)

(I) Faculty of Science and Technology,Sophia University
Mailing address:Department of Electrical and Electronic Engineering
Faculty of Science and Technology Sophia University
7-1,Kioicho,Chiyodaku,Tokyo,Japan
Telephone:03-3238-3421 Telefax:03-3238-3321

(II) Institute of Industrial Science, University of Tokyo
Mailing address:Institute of Industrial Science, University of Tokyo
7-22-1,Roppongi,Minatoku,Tokyo,Japan
Telephone:03-3402-6321 ex 2646 Telefax:03-3402-6226

'93 SPIE's International Symposium on Optical Engineering and Photonics in
Aerospace Science and Sensing
Proceedings of "Underground and Obscured Object Imaging and Detection"

Abstract

It has been developed the techniques which have been imaging optically opaque region using an electromagnetic wave radar in order to estimate the location of the objects in those region. One important application of these techniques is the detection of buried pipes and cables. In case of underground radar, its image quality often becomes low because the nature of soil is not uniform and an electromagnetic wave is attenuated in soil. Hence, the method which improves the quality of the radar images is required. Therefore, the synthetic aperture method for aircraft radar has been applied mainly. But it is a problem that this method requires a long computation time and the dielectric constant in soil have to be estimated, which is an essential parameter for this method. However, it is difficult to estimate the dielectric constant automatically. So it was done using human recognition. In this paper, we point out that the quality of underground images can be improved significantly by means of the f-k migration method, and propose a method of estimating dielectric constant in soil using the peak to peak value of processed images. Furthermore, we applied f-k migration method into the divided image blocks in order to take account of nonuniformity of dielectric constant in soil. In this method LOT(Lapped Orthogonal Transform) was applied. LOT is a new block transform method whose basis functions overlap in adjacent blocks, and it has a fast computation algorithm. The result of applying the block migration method into the underground radar images are presented it points out the good capability for the image quality improvement taking account of the dielectric constant of each block and the application of LOT can improve the processing time.