CAJ | 학술논문

以新疆特色水果哈密瓜为研究对象，提出一种利用 Matlab 辅助 DSP提取哈密瓜纹理的方法。该方法首先利用Matlab把待处理的图像转换为数据文件（ dat 文件），然后利用 DSP 强大的运算能力分割哈密瓜表面纹理特征，并将分割结果同样以dat文件传递给MatLab ，最后利用MatLab强大的数据分析与图像显示能力建立哈密瓜分类模型。实验结果表明，该方法能够有效地对哈密瓜表面纹理进行分割与分类，分类准确率为88．10％。该方法不但能够缩短DSP系统的开发周期，且能够为今后开发基于DSP的哈密瓜品质实时检测系统奠定基础。
이신강특색수과합밀과위연구대상，제출일충이용 Matlab 보조 DSP제취합밀과문리적방법。해방법수선이용Matlab파대처리적도상전환위수거문건（ dat 문건），연후이용 DSP 강대적운산능력분할합밀과표면문리특정，병장분할결과동양이dat문건전체급MatLab ，최후이용MatLab강대적수거분석여도상현시능력건립합밀과분류모형。실험결과표명，해방법능구유효지대합밀과표면문리진행분할여분류，분류준학솔위88．10％。해방법불단능구축단DSP계통적개발주기，차능구위금후개발기우DSP적합밀과품질실시검측계통전정기출。
In this study, Hami melon which is a characteristic fruit in Xinjiang was taken as the research object.The method of surface texture feature segmentation using Matlab and DSP was put forward.First, the original image was transformed to data file (*.dat) .Then, texture features of Hami melon was segmented using DSP, and the result was then imported into Matlab in the format of DAT .Finally, classification model of Hami melon was established with the function of data analysis and image display.The results showed that this method can effective classify Hami melon based on the surface texture feature segmentation.The accuracy was 88.10 %.The development period of DSP system was shortened, which will provide evidence for real-time detection system of Hami melon quality in the future.