CAJ | 학술논문

为了实现干制红枣的大小自动分级，介绍了应用机器视觉的干制红枣自动分级方法，利用 CCD 摄像机获取红枣的样本图像，应用MATLAB软件编程实现了样本图像的灰度化、二值化、图像分割、图像滤波、图像形态学处理、边缘检测和特征量提取等处理，参照干制红枣分级标准完成了红枣自动分级。通过实验数据回归分析得出红枣实测纵径、果质量与识别值当量值之间的数学检测模型，其决定系数分别达0．9955和0．9481。实验表明，采用句法模式识别对数据进行红枣大小分级，分级准确率达85％。
위료실현간제홍조적대소자동분급，개소료응용궤기시각적간제홍조자동분급방법，이용 CCD 섭상궤획취홍조적양본도상，응용MATLAB연건편정실현료양본도상적회도화、이치화、도상분할、도상려파、도상형태학처리、변연검측화특정량제취등처리，삼조간제홍조분급표준완성료홍조자동분급。통과실험수거회귀분석득출홍조실측종경、과질량여식별치당량치지간적수학검측모형，기결정계수분별체0．9955화0．9481。실험표명，채용구법모식식별대수거진행홍조대소분급，분급준학솔체85％。
In this paper , in order to realize the automatic classification of the size of the red jujube .The machine vision method was adopted in automatic red jujube hierarchy .The sample image was shot by a CCD camera and a series of treatments were proposed , with the soft of MATLAB , including graying , binary treatment , image segmentation , image fil-ter , image morphology processing , edge detection , the picking up of eigenvalues and so on .According to the standard of the red jujube hierarchy accomplished the automatic red jujube hierarchy .Through testing and analysis , we got the math-ematical model for the relationship between measured fruit longitudinal diameter , fruit weight and recognition fruit equiva-lent value , and the R-square is 0 .9955 and 0 .9481 .The date of red jujube ’ s size was graded by syntactic pattern rec-ognition .The results indicated that accuracy of classification was 85%.