This paper presents a fully automatic method to reconstruct the relative orientation of a pair of stereo images by matching area features. First, by using the concept of coarse-to­fine process, the stereo images are stored in a form of image pyramid. In the coarsest level, area features of homogeneous regions are extracted by using region-growing method. Then, their boundaries are described with Fourier descriptors. Applying the least-squares approach to matching Fourier descriptors, the recognition information consists of shape similarity and relative orientation between each possible matching pairs of area features is obtained. Based on the recognition information, conjugate area features can be matched by using Hopfield-Tank neural networks. When conjugate area features are determined, the approximate orientation between images can be easily solved. Knowing the approximate orientation, area-based matching, through a coarse-to-fine matching process, can be applied to obtain accurate tie points, and relative orientation can be solved. A pair of real aerial images were tested. The results show a successful application of the method.

Hough Transform is an effective method for extracting geometric primitives from image. However, the standard scheme suffers from the correlated error when the image content is complex. In this study, the orientation constrained Hough transform is proposed for the extraction of building edges from the scanned aerial photo. Based on the chain code concept, twelve templates are defined for the screening of line components to reduce the range of the parameter space. This procedure reduces the number of transformations required, and also limits the influence of the correlated error. Numerical experiments are performed with both the standard Hough transform and the proposed scheme. Three robustness indices, SNR, and the execution time are computed for evaluation.

ERS-1 SAR images were used to evaluate the directional wave spectra. The remote­sensing techeniq is applied and asists the traditional measuring method to get better direc­tional spectra easily. Two-dimensional F.F.T (Fast Fourier Transform) algorithm was applied to caculate two dimensional wave spectrum. To improve the resolution of directional wave spectra is the most impressing work in this study. The estimated SAR spectra were compared with those measured by wave-rider. From the comparison of SAR and in­situ spectra, the feasibility of estimating ocean wave directional spectrum from SAR images was confirmed in this study.

The geometric relationships between the SPOT stereo images and the ground points can be approximately described by simplified polynomials in place of collinearity equations. Because the SPOT image is obtained by the so-called "push broom" scanning principle the position of projecting centers is a time-dependent function. Using rigorous collinearity equations to solve space intersection problems is much more difficult than that of the aerial photography. Kratky proposed the polynomial mapping functions (PMFs) to solve the problem of real-time positioning in analytical plotter. Here, we attempt to apply them to the matching problems of digital SPOT stereo images. The PMFs in this research have three major applications：First, the polynomial func­tions for the image-to-image relationships are used to define geometric constraints in image space. Second, the image-to-object PMFs are used to perform space intersection to get the ground coordinates. Finally, the object-to-image PMFs are used to project grid points of the DTM up ward to SPOT images for orthoimage generation. The matching of SPOT stereo panchromatic images using PMFs can achieve an accu­racy of 10 m in elevation. Moreover, PMFs can be used to derive a very good approximation of the epipolar lines, which can be used to reduce the search space from 2-D to 1-D and also increase the success rate of the image matching.