【摘要】：近年來(lái),水果產(chǎn)業(yè)隨著鏈?zhǔn)缴a(chǎn)、網(wǎng)絡(luò)營(yíng)銷(xiāo)以及混合發(fā)展等發(fā)展模式,顯示出廣闊的市場(chǎng)前景。如何提高水果競(jìng)爭(zhēng)力,提升水果產(chǎn)品附加值,將是水果產(chǎn)業(yè)的發(fā)展重點(diǎn)。分揀技術(shù)作為水果商品化的重要途徑,對(duì)于提高水果品質(zhì),增加農(nóng)業(yè)現(xiàn)代化都有這十分重要的意義。水果分揀系統(tǒng)在一定程度上增加了水果產(chǎn)品的附加商品值,提高了水果產(chǎn)業(yè)的利潤(rùn)空間。對(duì)于分揀系統(tǒng),它的運(yùn)行效率和分揀準(zhǔn)確度尤為重要。本論文通過(guò)引入監(jiān)督學(xué)習(xí)的Adaboost算法改進(jìn)現(xiàn)有的Canny邊緣檢測(cè)算法,使分揀系統(tǒng)對(duì)臍橙邊緣的識(shí)別更加準(zhǔn)確,執(zhí)行效率更高。針對(duì)臍橙的圖像特征,從形狀特征、變換特征等方面對(duì)水果圖像進(jìn)行了分析和討論。類(lèi)比分析圖像特征與體積等的關(guān)系,對(duì)臍橙的體積進(jìn)行了公式估計(jì),并據(jù)此計(jì)算出臍橙的密度等數(shù)據(jù),為其分揀提供了更多的分揀標(biāo)準(zhǔn)。首先,本文闡述了課題的研究背景及水果分揀系統(tǒng)的國(guó)內(nèi)外研究成果及其意義,對(duì)圖像處理和圖像識(shí)別的基本理論和方法做了具體說(shuō)明。然后,文章簡(jiǎn)要介紹了圖像邊緣檢測(cè)算法,提升了該算法的抗噪能力,并通過(guò)Adaboost算法改進(jìn)邊緣連接,使圖像連接更符合人的認(rèn)識(shí)。對(duì)圖像特征進(jìn)行了分析和選取,并進(jìn)行實(shí)驗(yàn)分析,設(shè)計(jì)樣本臍橙的體積模型,完成對(duì)圖像中水果的擬合。接著,本文設(shè)計(jì)并搭建了完整的臍橙分揀系統(tǒng),編寫(xiě)了控制應(yīng)用軟件,簡(jiǎn)述了工作流程,并使整套系統(tǒng)正常高效地進(jìn)行臍橙的數(shù)據(jù)采集和分揀。最后,本文對(duì)工作進(jìn)行了全面的總結(jié),說(shuō)明了分揀系統(tǒng)的優(yōu)勢(shì)特點(diǎn),概括了全文的主要?jiǎng)?chuàng)新點(diǎn)和研究成果,針對(duì)系統(tǒng)的不足和今后的擴(kuò)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.
【學(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類(lèi)號(hào)】：TS255.3;TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 郝紅衛(wèi);王志彬;殷緒成;陳志強(qiáng);;分類(lèi)器的動(dòng)態(tài)選擇與循環(huán)集成方法[J];自動(dòng)化學(xué)報(bào);2011年11期

2 磨少清;劉正光;張軍;韋衛(wèi)星;;基于圖像自身信息的圖像邊緣檢測(cè)閾值自動(dòng)設(shè)定方法[J];光電子.激光;2011年08期

3 苑瑋琦;王楠;;基于局部灰度極小值的掌脈圖像分割方法[J];光電子.激光;2011年07期

4 唐路路;張啟燦;胡松;;一種自適應(yīng)閾值的Canny邊緣檢測(cè)算法[J];光電工程;2011年05期

5 曲迎東;李榮德;白彥華;李潤(rùn)霞;馬廣輝;;高速的9×9尺寸模板Zernike矩邊緣算子[J];光電子.激光;2010年11期

6 林開(kāi)顏;吳軍輝;;基于計(jì)算機(jī)視覺(jué)的水果分級(jí)技術(shù)研究進(jìn)展[J];信息化縱橫;2009年10期

7 何文浩;原魁;鄒偉;;自適應(yīng)閾值的邊緣檢測(cè)算法及其硬件實(shí)現(xiàn)[J];系統(tǒng)工程與電子技術(shù);2009年01期

8 范生宏;黃桂平;陳繼華;李廣云;周華;;Canny算子對(duì)人工標(biāo)志中心的亞像素精度定位[J];測(cè)繪科學(xué)技術(shù)學(xué)報(bào);2006年01期

9 應(yīng)義斌,饒秀勤,黃永林,王劍平;運(yùn)動(dòng)水果圖像的實(shí)時(shí)采集方法與系統(tǒng)研究[J];農(nóng)業(yè)機(jī)械學(xué)報(bào);2004年03期

10 沈明霞,李秀智,姬長(zhǎng)英;水果品質(zhì)檢測(cè)中的模糊閾值分割方法[J];農(nóng)業(yè)機(jī)械學(xué)報(bào);2003年05期

相關(guān)碩士學(xué)位論文 前2條

1 侯大軍;基于機(jī)器視覺(jué)的蘋(píng)果特征選擇和分類(lèi)識(shí)別系統(tǒng)[D];江蘇大學(xué);2010年

2 龐江偉;基于計(jì)算機(jī)視覺(jué)的臍橙表面常見(jiàn)缺陷種類(lèi)識(shí)別的研究[D];浙江大學(xué);2006年

，

本文編號(hào)：2468129