摘 要

基于机器视觉的苹果缺陷与特征的检测对于提高我国水果产后处理和市场销售综合竞争力具有相当重要的意义。我国是水果生产大国，但是自动化检测的水平一直低于世界发达国家，严重影响了我国水果的国际竞争力。本论文所讨论是利用机器视觉检测苹果的缺陷与特征，在充分参考国内外的发展研究成果之后，创新特色，合理选用合适的光源，摄像机，处理器等。利用机器视觉对苹果图像进行分析，通过对其纹理，特征值，颜色等方面的特征采取，来分辨水果的好坏，水果的成熟度以及水果的大小集一体的水果检测系统。在后期处理的过程中利用MATLAB来加快水果的检测，提高苹果检测的正确率。

首先对苹果的图像进行采集，读取所需要的苹果的图像。然后对苹果的图像进行平滑，分割，灰度化和锐化等图像处理，然后一次提取出苹果的大小，颜色和缺陷参数。运用MATLAB函数找出圆心然后利用直径判别法，计算出苹果的大小。在颜色的辨别方法中，，建立HIS颜色空间，根据颜色空间来判断苹果的成熟度，分级正确率不低于百分之九十，而缺陷分级采用梯度特征提取，通过梯度特征来判断苹果是否有缺陷。

关键词：机器视觉，图像处理，苹果分级，颜色，大小，缺陷

Apple defects based on machine vision and feature detection based on machine

ABSTRACT

Detection for the characteristics and defects of the apple based on machine vision has important significance to the improvement of fruit postpartum processing and marketing comprehensive competitiveness in our country. Our country has a big fruit production , but automation testing level has been lower than developed countries in the world, and seriously affect the international competitiveness of Chinese fruit. This paper discusses the use of machine vision detection of the defects and characteristics of apple, after full reference for the development of the research results at home and abroad, with the characteristics of innovation, reasonable choice of suitable light source, camera, processor and so on, by using machine vision for apple images is analyzed, the detection system based on its texture, characteristic values of color characteristics to take, to distinguish the stand or fall of fruit, fruit maturity and the size of the fruit . In the process of post processing using MATLAB to accelerate the detection of fruit, improve the apple test of the time. To be read by the first image acquisition, obtain apple's image. Then smooth apple's images, segmentation, gray and sharpening image processing, and then an extract of apple's size, color and defect parameters. Use MATLAB function to find the center of the circle and then using criterion in diameter, on the classification, the accuracy can reach above eighty percent. In the color discrimination method, the use of visual network as the classifier, establish HIS color space, judged by the color space of apple's maturity, the classification accuracy is not lower than ninety percent, feature extraction and defect classification using gradient, through gradient characteristics to determine whether apple is flawed.

Keywords: machine vision;image processing;the apple grading;color;size;defects

目录

目录..........................................................................1

1.绪论........................................................................3

1.1.课题研究背景及意义...................................................... .3

1.2.国内外同类研究发展趋势...................................................3

1.2.1.国外研究现状........................................................4

1.2.2国内研究现状............................... ..........................4

1.3.本研究内容...............................................................5

2.机器视觉硬件设计.............................................................7

2.1.机器视觉系统的概述......................................................7

2.2机器视觉系统硬件设计......................................................8

2.2.1光源设计.................... ................................. ....... 8

2.2.1．1衡量光源的好坏................................................8

2.2.1．2照明方式......................................................8

2.2.1．3光源选择......................................................9

2.3相机与镜头................................................................9

2.4传送带............................................................. ......10

2.5通过硬件系统采集的苹果图片...............................................11

3苹果图片的预处理......... ....................................................12

3.1图片平滑滤波处理.........................................................12

3.2.图像灰度化处理.........................................................15

3.3图像锐化处理.... .........................................................16

3.4.图像分割处理......... ...................................................17

3.5本章小结......... .......................................................19

4苹果大小的检测.. .............................................................20

5苹果颜色的检测.................... .......................... ................21

5.1基于HIS颜色模型的苹果成熟度的检测.......................................21

5.2颜色特征的提取...........................................................21