基于机器视觉的花生仁粒数检测的方法.doc

广西科技大学 毕业设计（论文）说明书 课题名称 花生仁粒数检测算法的研究（中文） Peanut kernels detection algorithm research （英文） 系 别 专 业 班 级 学 号 姓 名 指导教师 2012年 05 月 26 日 摘要 花生作为我国主要的经济作物之一，其主要用于生产植物油。能快速的估算出花生里的花生仁粒数所占的比例，对于预测一大批花生的产油量，提高企业的效用，很有必要。 本文基于机器视觉的花生仁粒数的检查，对花生选取合适的背景，在光照886LUX的条件下，采用黑色作为拍摄图像的背景，进行拍摄花生图像。再通过otsu法对花生图像二值化处理，估算花生图像的面积，通过最小二乘法对花生图像面积和花生颗粒的实际花生仁粒数进行拟合，从而达到估算出花生的花生仁粒数。再通过估算出来的花生仁粒数与花生颗粒的花生仁粒数进行比较，得出实验方法的误差精度。实验中，一共拍摄了花生图像338幅：其中出现一定误差的有42幅，算出实验的精度为12.426%。试验能达到预期的要求。 关键字：花生，图像，二值化，二乘法 Abstract Peanuts as one of Chinas major economic crops, mainly for the production of vegetable oil. Can quickly estimate the proportion of peanut in the peanut grains, and improving the effectiveness, it is necessary to predict a large number of peanut oil production. Based on the inspection of machine vision peanut grains, peanuts, select the appropriate background, under the in light 886LUX conditions, using black as the background of the captured images, to shoot the peanut image. Then the OTSU France peanut binary image processing to estimate the peanut image area were fitted by the least squares method peanut image area, and peanuts were peanut grains, so as to achieve the estimate peanut peanut grains. Estimated from the peanut grains and peanuts were peanut grains comparison, the error precision of the experimental method. Experiment, shooting a total of 338 peanut image: 42 in which a certain error, calculate the experimental accuracy of 12.426%. The test can meet expectations. Keywords: Peanut, image, Binarization，Two multiplication 目录 TOC \o 1-3 \h \z \u HYPERLINK \l _Toc326222320 1. 绪论 PAGEREF _Toc326222320 \h 1 HYPERLINK \l _Toc326222321 1.1 基