基于MatLab和DSP的哈密瓜纹理分割与分类

Texture Segmentation and Classification of Hami Melon Based on MatL ab and DSP

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Mat Lab,最后利用Mat Lab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。

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.