本文選題：白背飛虱測(cè)報(bào)　切入點(diǎn)：圖像處理　出處：《浙江理工大學(xué)》2017年碩士論文

【摘要】：稻飛虱是我國(guó)水稻上最重要的一類遠(yuǎn)距離遷飛性害蟲,各稻區(qū)以褐飛虱,白背飛虱和灰飛虱最為常見,它們常群居在水稻中下部取食或產(chǎn)卵為害。稻飛虱不同蟲態(tài)的田間發(fā)生調(diào)查是其進(jìn)行準(zhǔn)確測(cè)報(bào)和合理防治的關(guān)鍵。在大部分水稻種植國(guó)家,稻飛虱田間調(diào)查一般采用人工拍查法、目測(cè)法、掃網(wǎng)法和燈誘法,其中,應(yīng)用最為廣泛是拍查法。“拍查法”查獲的飛虱率受到蟲口密度、水稻生育期、盤內(nèi)壁濕潤(rùn)程度等影響且容易造成調(diào)查者身體和視覺疲勞,調(diào)查效率低下。人工田間調(diào)查時(shí)只記錄了稻飛虱種類和各種蟲態(tài)數(shù)量的數(shù)據(jù),后面無法追溯田間調(diào)查時(shí)飛虱發(fā)生的真實(shí)情況。劉慶杰等[1-4]充分利用數(shù)字圖像處理技術(shù),研究了不同特征對(duì)水稻基部稻田飛虱檢測(cè)的影響,取得了較好的檢測(cè)效果,但飛虱誤檢率仍舊偏高。本文在此基礎(chǔ)上,采用新的三層檢測(cè)方法,研究了圖像特征和分類模型參數(shù)的選擇對(duì)水稻基部飛虱檢測(cè)率和誤檢率的影響,以及白背飛虱各蟲態(tài)(包括長(zhǎng)翅型成蟲、短翅型成蟲、高齡若蟲和低齡若蟲)的分類問題。本論文主要研究?jī)?nèi)容、研究結(jié)果和創(chuàng)新點(diǎn)包括:(1)在第一層檢測(cè)中,研究了不同維數(shù)的HOG特征和不同級(jí)聯(lián)層數(shù)的Adaboost分類器對(duì)水稻基部白背飛虱檢測(cè)率和誤檢率的影響。首先,對(duì)2012-2016年期間采集的水稻基部飛虱圖像建立了白背飛虱和非飛虱噪聲的正負(fù)訓(xùn)練樣本集;然后,提取訓(xùn)練樣本不同維數(shù)的HOG特征;利用HOG特征訓(xùn)練不同級(jí)聯(lián)層數(shù)的Adaboost分類器,用于檢測(cè)稻飛虱;最終選擇最優(yōu)Adaboost分類器測(cè)試525張水稻基部飛虱圖像。結(jié)果表明該算法對(duì)白背飛虱檢測(cè)率為90.7%,誤檢率為56.2%。(2)在第二層檢測(cè)中,針對(duì)第一層中存在較多的誤檢噪聲,研究了不同局部圖像特征訓(xùn)練的SVM分類器對(duì)非飛虱噪聲識(shí)別情況。這些噪聲主要包括水珠、水面反光、泥點(diǎn)以及稻葉,它們?cè)诩y理上與飛虱存在較大差異。首先提取訓(xùn)練樣本的Gabor與LBP特征,使用Z-score進(jìn)行歸一化;然后,利用Gabor、LBP和兩個(gè)特征融合來訓(xùn)練SVM分類器,根據(jù)不同特征訓(xùn)練獲得的SVM分類器ROC曲線,發(fā)現(xiàn)Gabor和LBP融合的紋理特征訓(xùn)練的SVM分類器對(duì)白背飛虱和非飛虱噪聲識(shí)別率高;最終利用該SVM分類器對(duì)525張第一層檢測(cè)后得到的子圖像進(jìn)行非飛虱噪聲排除;結(jié)果表明該算法將第一層的誤檢率從56.2%降低到了10.2%。(3)在第三層檢測(cè)中,研究了白背飛虱不同蟲態(tài)分類識(shí)別的問題。針對(duì)不同蟲態(tài)的白背飛虱HOG特征差異顯著,本文首先提取了白背飛虱三種蟲態(tài)長(zhǎng)翅型成蟲、高齡若蟲和低齡若蟲HOG特征,并使用Z-score進(jìn)行歸一化;然后,利用PCA與LDA方法對(duì)HOG特征進(jìn)行降維,比較不同降維算法對(duì)白背飛虱不同蟲態(tài)的識(shí)別性能;最后,采用SVM分類器對(duì)525張水稻基部飛虱圖像第二層檢測(cè)到的白背飛虱子圖像進(jìn)行蟲態(tài)分類識(shí)別;結(jié)果表明該算法對(duì)白背飛虱長(zhǎng)翅型成蟲、高齡若蟲和低齡若蟲識(shí)別率分別為93.2%、82.7%和86.9%。綜合三層檢測(cè)結(jié)果,最終獲得水稻基部白背飛虱各蟲態(tài)平均識(shí)別率為73.1%。誤檢率為23.3%。對(duì)于無蟲的圖像,誤檢率為5.6%。由此可見,利用圖像處理方法進(jìn)行水稻基部飛虱測(cè)報(bào)是可行的。
[Abstract]:Rice is the most important rice in China on a long-distance migratory insect, the rice to brown planthopper, sogatellafurcifera and l.striatellus was the most common, they often populations in Rice under feeding or oviposition. Occurrence of rice planthopper infestation of different insect state investigation is the key to accurate forecasting and rational prevention in most countries. Planting rice, rice field investigation using artificial shoot check method, visual method, sweep net method and light trap method, among them, the most widely used is to take check method. "Shoot check method" seized by planthopper insect density, rice growing period, moist degree and the influence of wall plate easy to cause the body survey and visual fatigue investigation efficiency. Artificial field investigation only records the number of rice planthopper types and various stages of data, the real situation cannot be traced back behind the field investigation. Liu Qingjie planthopper occurrence [1-4] charge Using digital image processing technology, the effects of different characteristics on the base of rice brown planthopper in rice field detection, a good detection effect, but the planthopper false detection rate is still high. On this basis, using three layers of new detection methods, study the image features and classification of the model parameters on the effect of rice base planthopper detection rate and false detection rate, and the white backed planthopper of different stages (including the macropterous and brachypterous adults, aged nymphs and young nymphs) classification problems. The main contents of this dissertation, the research results and innovations include: (1) in the first layer on the HOG feature detection. Different dimensions and the associated number of Adaboost classifier on the white back planthopper detection rate and false detection rate. First of all, on the base of the rice planthopper image acquisition period of 2012-2016 years, established the white backed Planthopper and the noise is non planthopper The negative training set; then, HOG features are extracted from different dimensions of training samples; Adaboost classifier based on HOG features of different layers of the cascade training, for the detection of rice planthopper; finally choose optimal Adaboost classifier test 525 rice base planthopper images. The results show that the algorithm WBPH detection rate is 90.7%, the error rate is 56.2%. (2 in the second layer) detection, the false noise exists in the first layer, the SVM classifier trained on different local image features non noise recognition. These noise planthopper mainly include water, water surface reflection, mud and rice leaves, they in the texture and planthopper first extract the Gabor there is a big difference. With the LBP feature of training samples, then using Z-score normalization; using Gabor, LBP and two features to train SVM classifier, SVM classifier based on ROC features of the different training The curve, found that Gabor and LBP fusion texture feature training SVM classifier WBPH and non noise planthopper high recognition rate; finally uses the SVM method to get the first 525 layers after detection of sub images are non noise planthopper exclusion; the results show that the algorithm will first layer error rate is reduced from 56.2% to 10.2%. (3) in the third layer detection, study the WBPH different insect state classification problems. According to the difference of HOG features sogatellafurcifera different stages significantly, the paper extracts sogatellafurcifera three instars macropterous nymphs, elderly and young nymph HOG characteristics, and normalized using Z-score; then, to reduce the dimensionality of HOG features using PCA and LDA method, the recognition performance comparison of different dimensionality reduction algorithm of WBPH at different developmental stages; finally, using SVM method to detect 525 rice planthopper images of second Zhang base layer to the White - backed image classification and recognition of lice insect state; the results show that the algorithm to WBPH macropterous adults, aged nymphs and nymph recognition rates were 93.2%, 82.7% and three layer 86.9%. results comprehensive detection, finally get the white back planthopper insect the average recognition rate is 73.1%. error rate is 23.3%. for no image of the insect, the false detection rate is 5.6%. thus, using the image processing method of rice planthopper at base is feasible.

【學(xué)位授予單位】：浙江理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：S435.112.3;TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 YAO Qing;XIAN Ding-xiang;LIU Qing-jie;YANG Bao-jun;DIAO Guang-qiang;TANG Jian;;Automated Counting of Rice Planthoppers in Paddy Fields Based on Image Processing[J];Journal of Integrative Agriculture;2014年08期

2 江偉堅(jiān);郭躬德;賴智銘;;基于新Haar-like特征的Adaboost人臉檢測(cè)算法[J];山東大學(xué)學(xué)報(bào)(工學(xué)版);2014年02期

3 鄒修國(guó);丁為民;劉德營(yíng);趙三琴;;基于4種不變矩和BP神經(jīng)網(wǎng)絡(luò)的稻飛虱分類[J];農(nóng)業(yè)工程學(xué)報(bào);2013年18期

4 肖晶晶;金志鳳;徐紅星;李仁忠;;水稻稻飛虱遷飛的研究進(jìn)展[J];中國(guó)農(nóng)學(xué)通報(bào);2013年18期

5 鄒修國(guó);丁為民;劉德營(yíng);趙三琴;;基于改進(jìn)Hu矩和遺傳神經(jīng)網(wǎng)絡(luò)的稻飛虱識(shí)別系統(tǒng)[J];農(nóng)業(yè)機(jī)械學(xué)報(bào);2013年06期

6 鄒修國(guó);;基于計(jì)算機(jī)視覺的農(nóng)作物病蟲害識(shí)別研究現(xiàn)狀[J];計(jì)算機(jī)系統(tǒng)應(yīng)用;2011年06期

7 任_g;顧成成;;基于HOG特征和SVM的手勢(shì)識(shí)別[J];科技通報(bào);2011年02期

8 胡潔;;高維數(shù)據(jù)特征降維研究綜述[J];計(jì)算機(jī)應(yīng)用研究;2008年09期

9 王東;吳湘濱;;利用粒子群算法優(yōu)化SVM分類器的超參數(shù)[J];計(jì)算機(jī)應(yīng)用;2008年01期

10 張建偉;王永模;沈佐銳;;麥田蚜蟲自動(dòng)計(jì)數(shù)研究[J];農(nóng)業(yè)工程學(xué)報(bào);2006年09期

，

本文編號(hào)：1694867