very small aperture terminal是什么意思

In this paper, the research of automatic registration between Aperture Radar Synthetic (SAR) images and optical remote sensing images is carried out.. In view of the SIFT operator to heterologous image feature point sets are extracted with corresponding disadvantages, this paper presents a based on line feature of SAR image coarse registration algorithm. On this basis, combined with based on mutual information registration method, the realization of SAR image and optical image automatic registration accuracy. This paper has the features of accurate and robust registration of SAR images and optical images with the features of the images of the images with the features of the obvious linear features.. The registration method can provide enough image registration accuracy for SAR image change detection, dynamic monitoring, image fusion and target accurate positioning..
This study will focus on the main features of image registration based on line feature as the main feature of image registration.. According to the difference between SAR images and optical images, the edge of different images is detected by different detection operators.. First, because of the SAR images are affected by multiplicative speckle severely affected, by the roewa edge detection operator detection in SAR image edge and at the same time optical image with high signal-to-noise ratio using traditional Canny operator to detect the optical image edge. And based on detected edge image by Hough transform algorithm characteristics of linear feature extraction, into the then the straight line between the plane geometry of the extracted linear feature as the feature point that is based on the point of intersection of a straight line as the candidate matching feature point set, after the use RANSACK points outside the removal and obtain optimal corresponding feature points and thus the coarse registration. Finally, the blocks to compute the mutual information of the image, further refining the same feature, to achieve high precision registration of coarse registration of SAR images. The registration results of the measured data proved the validity of this method..