水果表面缺陷自动检测系统中的人工智能方法研究

通过计算机视觉技术获取了带有各种表面缺陷的苹果图像并进行预处理，采用自适应特征聚类（ＳＡＦＣ）神经网络与模糊加权决策树（ＦＷＤＴ）相结合的方法实现了缺陷区域的准确检测和详细分类。实验结果表明，用人工智能方法进行表面缺陷检测，具有良好的抗噪容错能力并能有效地克服传统图像分割方法适应性差的缺点，提高判别准确率和分类精度。

Study on Artificial Intellectual Methods Used in Automated Detection for Fruit Surface Defect

The images of apples with different kinds of defects were acquired and preprocessed with the computer vision technology.Using the Self Adaptive Feature Clustering (SAFC) neural network and Fuzzy Weighted Decision Tree (FWDT) methods,the accurate detection and detailed classification of defective areas on apples were successively achieved.Experimental result showed that,in the detection of surface defects,artificial intellectual method has good anti noise and fault tolerance ability.It can effectively overcome the shortcomings of low adaptability of traditional image segmentation method,and thus improve the accuracy of defect detection and classification.