The Post-processing Of The Image For The Forward Looking Sonar

It is difficult and significant to detect and classify a small target under the sea. The contents in the dissertation originate from the controlling and displaying system of a sort of forward looking sonar. The sonar datum analysis, sonar image display, and image post-processing/recognition are included in the dissertation. The major contents are as follows in general.(1)the author gives the progress trends in image processing and current status of research on sonar image processing.(2)The sonar datum is analyzed, and the sonar image is reconstructed according to it. The controlling and displaying software concerned with datum acquisition is compiled.(3)The method of histogram equalization is attempted to used in the sonar image, and the result is analyzed. The author thoroughly analyzes the essence of a sort of algorithms on image enhancement with fuzzy sets and applies it to the sonar image enhancement.(4)The results of noise reduction with a median filter, a average filter and a gauss filter are analyzed. The author applies these algorithms to the sonar image smoothing.(5)The results of segmentation with the threshold estimated by iteration and the multiple thresholds are analyzed. These algorithms are used in the sonar image segmentation.(6)The algorithms of erosion, dilation and opening in mathematical morphology are analyzed. They are used inthe sonar image restoration.(7)The algorithm of template matching is analyzed, and they are used in the sonar image matching.(8)The theory of the image invariant moment is studied. Theseven invariant features of an image are tested by the emulational experiments, and the author analyzes the errors. They are also applied to the sonar image recognition.It is shown that the algorithms and approaches providing in the thesis are effective and applicable, and can be used for reference of the research of sonar image detection and recognition.