基于边缘检测改进算法的脐橙分拣系统设计与实现

发布时间：2019-04-29 08:50

【摘要】：近年来,水果产业随着链式生产、网络营销以及混合发展等发展模式,显示出广阔的市场前景。如何提高水果竞争力,提升水果产品附加值,将是水果产业的发展重点。分拣技术作为水果商品化的重要途径,对于提高水果品质,增加农业现代化都有这十分重要的意义。水果分拣系统在一定程度上增加了水果产品的附加商品值,提高了水果产业的利润空间。对于分拣系统,它的运行效率和分拣准确度尤为重要。本论文通过引入监督学习的Adaboost算法改进现有的Canny边缘检测算法,使分拣系统对脐橙边缘的识别更加准确,执行效率更高。针对脐橙的图像特征,从形状特征、变换特征等方面对水果图像进行了分析和讨论。类比分析图像特征与体积等的关系,对脐橙的体积进行了公式估计,并据此计算出脐橙的密度等数据,为其分拣提供了更多的分拣标准。首先,本文阐述了课题的研究背景及水果分拣系统的国内外研究成果及其意义,对图像处理和图像识别的基本理论和方法做了具体说明。然后,文章简要介绍了图像边缘检测算法,提升了该算法的抗噪能力,并通过Adaboost算法改进边缘连接,使图像连接更符合人的认识。对图像特征进行了分析和选取,并进行实验分析,设计样本脐橙的体积模型,完成对图像中水果的拟合。接着,本文设计并搭建了完整的脐橙分拣系统,编写了控制应用软件,简述了工作流程,并使整套系统正常高效地进行脐橙的数据采集和分拣。最后,本文对工作进行了全面的总结,说明了分拣系统的优势特点,概括了全文的主要创新点和研究成果,针对系统的不足和今后的扩展指明了方向。
[Abstract]:In recent years, with the development of chain production, network marketing and mixed development, fruit industry shows a broad market prospect. How to improve the competitiveness of fruit, enhance the added value of fruit products, will be the focus of fruit industry development. As an important way to commercialize fruit, sorting technology is very important for improving fruit quality and increasing agricultural modernization. The fruit sorting system increases the added value of fruit products to a certain extent, and improves the profit space of fruit industry. For sorting system, its operation efficiency and sorting accuracy are particularly important. In this paper, the supervised learning Adaboost algorithm is introduced to improve the existing Canny edge detection algorithm, so that the sorting system can recognize the navel orange edge more accurately and efficiently. According to the image features of navel orange, the fruit images were analyzed and discussed from the aspects of shape features, transformation features and so on. By analogical analysis of the relationship between image features and volume, the formula is used to estimate the volume of navel orange, and the density and other data of navel orange are calculated according to the formula, which provides more criteria for the sorting of navel orange. First of all, this paper describes the research background, fruit sorting system at home and abroad research achievements and significance, image processing and image recognition of the basic theory and methods are explained in detail. Then, this paper briefly introduces the image edge detection algorithm, improves the anti-noise ability of the algorithm, and improves the edge connection through the Adaboost algorithm to make the image connection more in line with people's understanding. The characteristics of the image were analyzed and selected, and the volume model of the sample navel orange was designed to fit the fruit in the image. Then, this paper designs and builds a complete navel orange sorting system, compiles the control application software, briefly describes the work flow, and makes the whole system to collect and sort the navel orange data normally and efficiently. Finally, this paper makes a comprehensive summary of the work, explains the advantages and characteristics of the sorting system, summarizes the main innovations and research results of the full text, and points out the direction for the deficiency of the system and the expansion of the system in the future.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TS255.3;TP391.41

【参考文献】