CAJ | 학술논문

为证实灰度信息在声呐图像中的重要价值，并获得有利于声呐图像目标分割的最佳增益值，研究了不同声呐增益值下，鲟鱼声呐图像灰度分布的形态差异，并基于灰度直方图分布信息的最大熵阈值分割算法对俄罗斯鲟（ Acipenser gueldenstaedti Brandt ）目标进行了分割，根据分割结果得到了最佳声呐增益值，即Gain＝15 dB。充分说明了在较优的声呐增益值下获取的声呐图像，即使目标形状发生严重变形时，利用其灰度直方图分布信息也能进行很好的目标分割，从而为解决鲟鱼声呐图像目标识别以及跟踪提供参考依据。
위증실회도신식재성눌도상중적중요개치，병획득유리우성눌도상목표분할적최가증익치，연구료불동성눌증익치하，심어성눌도상회도분포적형태차이，병기우회도직방도분포신식적최대적역치분할산법대아라사심（ Acipenser gueldenstaedti Brandt ）목표진행료분할，근거분할결과득도료최가성눌증익치，즉Gain＝15 dB。충분설명료재교우적성눌증익치하획취적성눌도상，즉사목표형상발생엄중변형시，이용기회도직방도분포신식야능진행흔호적목표분할，종이위해결심어성눌도상목표식별이급근종제공삼고의거。
In this paper, to prove that the gray level is the important information of the acoustic image, and obtain the best Gain value which is useful to partition the fish target, the effect of different gain values on gray level distribution of the acoustic images of Russian sturgeon( Acipenser gueldenstaedti Brandt) was studied, and the fish target was partitioned based on the gray level histogram distribution. The result showed that the best Gain value is 15 dB, under this Gain level, the fish target can be partitioned nicely from the acoustic image even when there is a serious target deformation. The work will provide useful information for sturgeon sonar image recognition and track.