KNN算法奶牛图像的奶牛身份识别研究+源代码
时间:2019-04-16 20:32 来源:毕业论文 作者:毕业论文 点击:次
| 摘要：为了方便奶牛场的管理人员对奶牛进行管理，本文以奶牛生活视频为对象进行基于机器视觉的奶牛身份识别。首先从视频中通过截取不同奶牛的4种方位（正面、背面、左侧、右侧）图像，创建了1294幅奶牛图像的图片集。然后使用图像增强、去噪、形态学重建等预处理方法和固定阈值分割的方法，得到了奶牛目标图像。最后提取奶牛图像的对比度特征、HSV特征、LBP纹理特征和LBPV纹理特征，运用KNN算法和支持向量机计算和比较不同特征下奶牛身份的识别率。本研究表明，对分割后的奶牛图像提取LBP特征和LBPV纹理特征并通过支持向量机的方法进行识别，识别率可以达到77.59%左右。34611 毕业论文关键词：奶牛身份识别；图像处理；特征提取；LBPV；支持向量机 Research on identification of cow based on cow image Abstract: In order to facilitate the manager of livestock farm to manage cows, this essay took cow life video as the object and the identification of cows was carried out based on machine vision. First of all, the video images of 4 positions (front, back, left and right) of different cows were intercepted, an image library containing 1294 cow images was established. Then, the method of image enhancement, noise reduction, morphological reconstruction and the method of fixed threshold segmentation were used to obtain the target of cow images. Finally, the contrast characteristics, HSV features, LBP texture features and LBPV texture features of cow images were extracted. The recognition rate of cows under different characteristics was calculated and compared by using KNN algorithm and support vector machines. The research demonstrates that LBP features and LBPV texture features were extracted from segmented cow images and identified by support vector machines, and the recognition rate can reach about 77.59%. Key words: cow identification; image processing; feature extraction; LBPV; support vector machine 目 录 摘要 1 关键词 1 Abstract 1 Key words 1 1 绪论 1 1.1 研究意义 1 1.2 国内外研究现状 2 1.2.1 国外研究现状 2 1.2.2 国内研究现状 2 1.3 研究方法 3 1.4 技术路线 3 2 奶牛图像库的建立 3 2.1 获取奶牛图像 3 2.2 人工识别奶牛身份 4 2.3 建立训练库与测试库 4 3 奶牛图像的预处理 5 3.1 灰度直方图 5 3.1.1 直接灰度变换 5 3.1.2 直方图均衡 5 3.2 图像平滑 6 3.2.1 均值滤波 6 3.2.2 中值滤波 7 3.3 形态学操作 7 3.3.1 灰度腐蚀 7 3.3.2 灰度膨胀 8 3.3.3 灰度开运算 8 3.3.4 灰度闭运算 8 3.3.5 形态学重建 8 4 奶牛图像的分割 8 4.1 边缘提取 9 4.1.1 拉普拉斯算子 10 4.1.2 Sobel算子 10 4.1.3 Roberts算子 11 4.1.4 Prewitt算子 11 4.2 阈值分割 12 4.2.1 迭代阈值分割 13 4.2.2 固定阈值分割 13 5 奶牛图像特征提取 16 5.1 图像的对比度特征提取 16 5.2 图像的颜色特征提取 16 5.3 基于LBP算法的纹理特征提取 17 (责任编辑：qin) |