Processing Of Composite Radar Images

The superposition display methods of electronic nautical chart and radar information can be grouped into three types, namely, the radar tracking target information display, the radar original video signal superposition display and the radar composite video image superposition display. This paper discusses the third approach to superimpose radar images and charts. The radar composite video image contains not only the original radar echo signals, but also EBL(Electronic Bearing Lines), range markers and the image information relating to radar operation menu. A way must be found to detect and erase these extra components in the composite image. The Hough transform is used to accomplish the detection of lines and circles in the composite radar image.A simple introduction to the superposition display of the radar image and the chart and popular Hough transform algorithms is made first. Then, an improved Hough transform algorithm suitable for the composite radar video image is proposed. This thesis focuses on different solution according to different combination of pixel points which Hough transform is interested in. The determination of point interval is a difficult part of this thesis. The utilization of "disjoin-join" idea is a key of this thesis, from which the algorithms of four-division, eight-division and nine-division are derived. "Join" processing or "disjoin" processing are two different solutions. The order of EBL detection and range markers detection is also discussed.In the process of erasing the signs of the radar video image, such as EBL and range markers, the method of retaining the radar echo signals is determined. A method based on the projection in the direction of distance is proposed for the erasing of range markers. The character recognition for the detection of such information as radar display range and bearing reference which are necessary to the superposition display of radar image and chart is briefly introduced. The template matching method is used for the recognition of the characters. In accordance with the above discussed approaches the corresponding algorithms are studied and programs are designed and realized. The converting from gray-level image to binary image and the template matching programs are written in MATLAB while the other parts of the software are developed with VC + + programming platform. Finally, some experiments are carried out. The results of the experiments show that the software can achieve the expected result and can provide a foundation for further study.