基于机器视觉技术的蔬菜新鲜度检测研究+源代码
毕业论文关键词：图像分割；特征提取；支持向量机；卷积神经网络；
Research on the freshness of vegetables based on machine vision technology
Abstract：In order to realize spinach freshness real-time inspection, this paper takes spinach as research target, adopts machine vision method to make freshness inspection. Firstly, through comparing the iterative threshold value choice method, of threshold segmentation values, OTSU method, regional growing and other algorithm, it adopts regional growing algorithm to part spinach image after morphological optimization, then extracts forms, textures, colors, GIST 13 features. Finally, by virtue of support vector machine (SVM) and Convolutional Neural Network (CNN) respectively it draws the characteristics to establish model and then presses ahead cross validation and identification inspection. Finally, with the characteristic of part binary system mode and application of support vector machine in building a recognition mode, the accuracy rate for recognizing a front lamina reaches 81.432%. the accuracy rate for recognizing a back lamina reaches 82.311%.
Key words:Image segmentation;Feature extraction;SVM;CNN
目 录
摘要 1
关键词 1
Abstract. 1
Key words 1
1绪论 1
1.1研究背景 2
1.2国内外研究现状 2
1.2.1国内研究现状 2
1.2.2国外研究现状 2
1.3研究的主要内容 3
2数据集的建立 4
2.1采集标准 4
3图像分割算法对比与选取 4
3.1阈值分割算法 4
3.1.1最大类间方差法 4
3.1.3迭代选择阈值法 6
3.2.1区域生长算法 7
3.2.3区域生长算法的优化 8
4特征提取 10
4.1形状特征 10
4.2颜色特征 10
4.3局部二进制模式 11
4.4GIST特征 12
5分类算法 12
5.1支持向量机 12
5.2卷积神经网络 12
6分类算法对比 14
6.1颜色分量+libsvm 14
6.1.1正面实验结果 14
6.1.2背面实验结果 15
6.2形状分量+libsvm 17
6.2.1正面实验结果 17
6.2.2背面实验结果 18
6.3LBP+libsvm 19
6.3.1正面实验结果 19
6.3.2背面实验结果 20
6.4GIST+libsvm 21
6.4.1正面实验结果 21
6.4.2背面实验结果 22
6.5卷积神经网络 22
6.5.1正面实验结果 23
6.5.2背面实验结果 23
7界面设计 24
8总结 25
致谢 25
参考文献 25
上一篇：基于专利地图的中国除草剂行业研究