基于机器视觉的圣女果分级分选机

发布时间：2018-01-30 18:05

本文关键词： 圣女果机器视觉分级分选图像处理嵌入式　出处：《宁夏大学》2017年硕士论文　论文类型：学位论文

【摘要】：圣女果因其营养价值高、口味清甜,深受人们的喜欢,具有巨大的市场潜力。完全依靠手工分类,不仅劳动强度大、生产效率低,且易造成果皮损失,影响美观和存储。因此,需要研究圣女果自动分级系统以降低人力成本,增加分选效率。为保证分选品质,本文以圣女果的颜色、大小、椭圆拟合程度等为标准对其进行分级,保证分级准确、快速、无损。具体主要完成了以下几项工作:1.本文广泛查阅国内外有关水果分级的资料,基于对机器视觉的研究,确定了适合圣女果分级分选机的整体机械设计方案、硬件选取方案及软件设计方案。2.搭建了完整的圣女果分级分选机结构,包括样本传输部分及控制观察箱部分。完成了设备所有零部件的设计,根据实际情况设计挡板结构和输送结构以保障圣女果分级过程的安全可靠。3.在图像预处理阶段,采用HSV颜色模型以增强抗干扰能力及识别的有效性,使其更易于维护与开发。由于圣女果自身的颜色与背景色差别较大,本文将采取阈值法对图像的色调值进行分割处理,从而达到去除背景的目的,依据最终效果以及所消耗的时间进行分析,使用形态学算法对圣女果图像去噪。4.依据采集到的圣女果图像特征,使用周长法判断其所属大小等级。根据图像所拥有的缺陷以及畸形程度,采用阈值法以及椭圆拟合算法进行识别以剔除不合格果,从而完成对圣女果的分选过程。通过最终试验结果表明,本文所使用的算法对圣女果进行分级分选的准确率较高且分选速度较快,拥有较高的实用价值以及研究意义。
[Abstract]:Because of its high nutritional value, sweet taste, deeply liked by people, it has great market potential. Totally relying on manual classification, it is not only labor intensity, low production efficiency, but also easy to cause peel loss. Therefore, it is necessary to study the automatic grading system to reduce the labor cost and increase the sorting efficiency. In order to ensure the quality of the sorting, this paper uses the color and size of the fruit. The ellipse fitting degree is used as the standard to classify it to ensure that the classification is accurate, fast and non-destructive. The following work has been done: 1.This paper extensively consulted the domestic and foreign information about fruit grading. Based on the research of machine vision, the overall mechanical design scheme, hardware selection scheme and software design scheme of the grade-sorting machine are determined. 2. The complete structure of the grade-sorting machine is built. Including sample transmission part and control observation box part. Completed the design of all parts of the equipment. According to the actual situation, the baffle structure and the conveying structure are designed to ensure the safety and reliability of the classification process. 3. In the stage of image preprocessing, the HSV color model is adopted to enhance the anti-interference ability and the effectiveness of recognition. Because of the difference between color and background color, this paper will take the threshold method to segment the color of the image, so as to get rid of the background. According to the final effect and the time consumed to analyze the use of morphological algorithm to the image denoising. 4. According to the collected characteristics of the holy fruit image. According to the defect and deformity degree of the image, the threshold method and ellipse fitting algorithm are used to identify the unqualified fruit. The final experimental results show that the algorithm used in this paper has higher accuracy and faster sorting speed. It has higher practical value and research significance.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S226.5;TP391.41

【相似文献】