本文關鍵詞：棉蚜發(fā)生量信息快速獲取方法與監(jiān)測模型的建立研究　出處：《石河子大學》2017年博士論文　論文類型：學位論文

  更多相關文章： 棉蚜 快速 監(jiān)測預警 模型建立

【摘要】：棉蚜發(fā)生量信息的快速、精準獲取是其科學精確防治的必要前提,而目前的棉蚜發(fā)生量信息采集依人工調(diào)查為主,受棉蚜分布不均勻和大量遷飛的影響,即使耗費大量的人力和時間,其數(shù)據(jù)調(diào)查精度和時效性仍然較差,致使棉蚜防治受到很大影響。因此,本研究針對目前棉蚜信息監(jiān)測依靠田間人工調(diào)查的方法,而導致棉蚜信息獲取滯后、采集圖像存在背景復雜和粘連等一系列問題。以棉蚜極強的趨黃特性為理論依據(jù),利用黃色粘蟲板在棉田進行大田試驗,通過對比分析不同高度和方向的黃色粘蟲板誘蟲效果,同時在人工調(diào)查對應小區(qū)不同類型棉蚜信息的基礎上,確定棉蚜信息最佳監(jiān)測條件。進而結合網(wǎng)絡高清拍照、無線遠程傳輸、機械自動化控制等現(xiàn)代信息技術,自主研發(fā)棉蚜發(fā)生量信息快速采集裝置,實現(xiàn)棉蚜圖像信息的自動采集,解決了采集圖像的背景復雜與蚜蟲粘連問題;進一步結合圖像識別計數(shù)技術構建棉蚜發(fā)生量信息快速獲取方法,在此基礎上建立棉蚜信息快速監(jiān)測預警模型,并開發(fā)了棉蚜信息快速監(jiān)測預警與決策一體化系統(tǒng),為農(nóng)業(yè)一線生產(chǎn)者提供所在地區(qū)棉蚜發(fā)生量信息查詢服務,同時為棉農(nóng)大田實際生產(chǎn)提供棉蚜防治決策信息。主要研究結果如下:1.基于趨黃性棉蚜最佳快速監(jiān)測條件的確定本研究基于棉蚜趨黃性進行大田試驗,通過設置不同方向和高度的誘蚜黃板進行誘蚜效果的統(tǒng)計分析,以此確定棉蚜信息最佳監(jiān)測條件。結果表明:棉田有翅蚜發(fā)生量和總蚜量之間的相關性極顯著,通過監(jiān)測有翅蚜的發(fā)生量可以很好的估測棉田總蚜量的發(fā)展趨勢;同時,不同高度和方向的黃板誘蚜量之間均存在極顯著的差異,并且通過兩因素方差分析,表明黃板高度和方向作為單一因子和交互作用均對誘蚜量的影響達到了極顯著水平,因此,黃板下邊沿距離棉花冠層的高度和黃板的方向?qū)γ扪列畔⒌谋O(jiān)測具有較大的影響。經(jīng)過2013年和2014年兩年試驗,結果證明,誘蚜黃板距離棉花冠層90cm高度且水平方向與60cm高度且向東方向的監(jiān)測黃板(即距離棉花冠層60cm-90cm),這兩個條件可以作為棉蚜發(fā)生量快速監(jiān)測的最佳條件。2.棉蚜發(fā)生量快速獲取方法的研究在棉蚜最佳監(jiān)測條件確定的基礎上,利用網(wǎng)絡高清拍照、無線遠程傳輸、機械自動化控制等現(xiàn)代信息技術,自主研發(fā)了棉蚜發(fā)生量信息快速采集裝置,實現(xiàn)了采集圖像的背景單一化,解決了蚜蟲堆積導致棉蚜識別計數(shù)精度不高的問題。通過對裝置進行測試試驗,結果表明,該裝置獲取圖片信息的有效性為95.83%、傳輸成功率達到99.4%、每次傳輸平均耗時為191s,其圖片信息的平均傳輸時間不影響棉蚜快速監(jiān)測預警的時效性。在此基礎上,基于圖像識別計數(shù)技術,以實用性、可擴充性、統(tǒng)一性和簡單性為設計原則,設計開發(fā)了棉蚜識別計數(shù)軟件,該軟件系統(tǒng)實現(xiàn)了棉蚜預警參數(shù)信息的自動采集和統(tǒng)計分析等功能,經(jīng)過人工計數(shù)對比測試分析,該系統(tǒng)識別計數(shù)的相對誤差為2.67%,完全滿足農(nóng)業(yè)生產(chǎn)中棉蚜預警決策的需求。3.棉蚜發(fā)生量實時估測模型的建立與驗證基于棉蚜發(fā)生量快速獲取裝置,于2015年進行了大田測試試驗,同時人工調(diào)查了棉田棉蚜實際發(fā)生量,通過分析裝置獲取信息和大田實際發(fā)生量進行相關關系,并在此基礎上建立了不同類型百株蚜發(fā)生量快速監(jiān)測模型。經(jīng)過2016年的模型驗證試驗表明:百株有翅蚜預測精度偏低,百株無翅蚜和百株全蚜量的預測精度較高。大田棉蚜爆發(fā)量的判斷一般依據(jù)全蚜量的發(fā)生量,并且百株全蚜量的預測值和真實值的均方根誤差(RMSE)僅為2180,因此,模型Y=0.0362x2-12.642x+16470對大田實際百株蚜發(fā)生量進行預測具有較好的可行性。與此同時,經(jīng)過大田實際調(diào)查對棉蚜危害程度進行了分級,并與基于棉蚜快速監(jiān)測裝置的百株蚜估測分級進行了對比分析,結果表明其預測等級完全一致,正確率達到100%。因此,本研究建立的棉蚜發(fā)生量預警模型完全可以滿足棉蚜發(fā)生程度的預警預測,再參考結合當?shù)貧庀筚Y料,能為大田棉蚜防治提供可靠的數(shù)據(jù)支持。4.棉蚜信息預警系統(tǒng)的研發(fā)本研究集棉蚜信息動態(tài)采集與圖像識別計數(shù)技術,結合前文建立的棉蚜快速監(jiān)測預警模型,集成計算機編程和數(shù)據(jù)庫技術,開發(fā)出棉蚜信息快速監(jiān)測預警與決策一體化系統(tǒng),系統(tǒng)主要包括信息展示、系統(tǒng)管理、信息管理、棉蚜預警決策,信息查詢與效益評價五個功能模塊。實現(xiàn)了通過對棉田棉蚜發(fā)生量的估測計數(shù),參考棉蚜發(fā)生程度和氣象資料提出預警信息,以此提出對應的防治措施,為棉農(nóng)提供準確的棉蚜防治時期和農(nóng)藥施用量,有效抑制棉花蟲害的大面積發(fā)生,提高了對棉蚜測報的準確率和時效性,同時節(jié)省了農(nóng)藥施用量,達到節(jié)本增效的目的。
[Abstract]:The amount of cotton aphid occurrence information fast, accurate access is a necessary precondition for the scientific and accurate control, and the amount of information collected by the artificial cotton aphid occurrence survey, by the impact of uneven distribution of cotton aphid and a large number of migration, even spend a lot of time and manpower, the investigation is still poor data accuracy and timeliness, resulting in the cotton aphid prevention by great influence. Therefore, this study focuses on the information monitoring on field investigation of artificial cotton, cotton aphid caused information lag, a series of problems with complicated background image. Adhesion to the cotton aphid strong trend yellow characteristics as the theoretical basis, a field experiment was conducted in the cotton field by yellow sticky board, yellow sticky board through the comparative analysis of different height and direction of the trapping effect, at the same time in the corresponding cell based artificial investigation of different types of cotton aphid information, determine the best monitoring information of cotton aphid Measuring conditions. And then combined with the network high-definition camera, wireless remote transmission, automatic control and other modern information technology, independent research and development of rapid information acquisition device of cotton aphid, Aphis gossypii to realize the automatic acquisition of image information, to solve the complex background image acquisition and aphid adhesion problems; combining with the construction of cotton aphid occurrence fast acquisition method of image information recognition count technology, on the basis of the establishment of rapid detection and early warning model information and the development of the cotton aphid Aphis gossypii, information rapid monitoring and early warning and decision system for agricultural producers, a region is provided for aphid information query service, at the same time as farmers field production cotton aphid prevention decision information. The main results are as follows: 1. to determine the trend of yellow the best conditions for the rapid monitoring of Aphis gossypii based on field experiment was carried out based on the trend of yellow, by setting the Statistical analysis of aphid trap effect in the same direction and height of the Yellow aphid, Aphis gossypii in order to determine the optimum conditions of information monitoring. The results show that the correlation between the occurrence of cotton aphids and aphid significantly, by monitoring the amount of aphids can estimate the total cotton aphid the good development trend; at the same time, different height and the direction of the Yellow induced significant differences were found between the amount of aphids, and through two factor variance analysis showed that the yellow, height and direction as the single factor and the interaction of impact induced aphid reached a very significant level, therefore, has a great influence on the direction of the edge information monitoring cotton cotton canopy height distance the yellow and yellow. After 2013 and 2014 two test results show that monitoring of aphid with yellow cotton canopy height and horizontal distance of 90cm direction and 60cm direction and height to the East Yellow (distance of cotton canopy, the best conditions for.2. 60cm-90cm) these two conditions can be used as a cotton aphid Aphis gossypii quantity rapid monitoring the occurrence of the amount of fast acquisition method based on determining the optimal monitoring conditions of the cotton aphid, the use of network high-definition camera, wireless remote transmission, automatic control and other modern information technology, independent research and development of cotton aphid fast acquisition device information, the realization of the background of a single image, resulting in the accumulation of cotton aphid to solve the aphid recognition and counting accuracy problem. Through the test, the results show that the effectiveness of the device, the device obtains image information is 95.83%, the transmission success rate of 99.4%, average time for the transmission of each 191s, the average the transmission time the picture information does not affect the timeliness of Aphis rapid monitoring and warning. On this basis, image recognition and counting technology based on practicality, can Extensibility, unity and simplicity of design principles, design and development of the cotton aphid counting software, the software system realizes the automatic data collection and statistical analysis functions of cotton aphid warning parameter information, through the manual counting contrast test and analysis, the system recognition and counting of relative error is 2.67%, fully meet the rapid establishment and verification of real time estimation the amount of acquisition device model based on the amount of cotton aphid Aphis gossypii warning decision in agricultural production needs of.3. cotton aphid were conducted in 2015, and investigated the test, artificial cotton aphid and the actual amount, through the analysis of means of access to information and the actual amount of field correlation is established based on the amount of rapid monitoring model different types of 100 plant aphid. Through the model test in 2016 showed that aphids per plant low prediction accuracy, 100 and 100 strains of apterae aphid strains The prediction accuracy is high. The amount of cotton aphid outbreak field based on the general judgment amount of full amount of aphids, and 100 strains of aphid prediction value and the real value of the root mean square error (RMSE) is only 2180, so the Y=0.0362x2-12.642x+16470 model in the field of actual amount per plant aphid occurrence forecast is feasible. At the same time after field survey of cotton aphid, harm degree is classified, and compared with the rapid monitoring device of 100 plant aphid Aphis gossypii estimation based on the classification, the results show that the prediction level is completely consistent, the correct rate of 100%. because of this, this study established a cotton aphides early-warning model can fully meet the prediction of the occurrence degree of cotton aphid then, the reference combined with local meteorological data, research can provide reliable data to support.4. information early warning system for the field of cotton aphid Aphis gossypii control in the study of cotton aphid information dynamic recovery Set and the image recognition and counting technology, combined with the rapid detection and early warning model previously established cotton aphid, integrated computer programming and database technology, the development of cotton aphid information fast monitoring and early warning and decision system, including system information display, system management, information management, Aphis warning decision five function module evaluation information query and achieve benefits. Based on the cotton aphid occurrence quantity estimation of reference counting, the occurrence degree of Aphis gossypii Glover and meteorological data provides early warning information, put forward the corresponding prevention and control measures, provide the period and amount of application of pesticide control of Aphis accurately for farmers, a large area of effective suppression of cotton pests and improve the forecasting accuracy of cotton aphid and timeliness, at the same time save the amount of pesticide application, to reduce cost and increase benefit.

【學位授予單位】：石河子大學
【學位級別】：博士
【學位授予年份】：2017
【分類號】：S435.622.1;S126

【相似文獻】

相關期刊論文 前7條

1 楊麗萍;隋學艷;楊潔;郭洪海;張錫金;黃玲;;山東省春季土壤墑情遙感監(jiān)測模型構建[J];山東農(nóng)業(yè)科學;2009年05期

2 段麗潔;廖玉芳;唐杰;卿湘濤;曾向紅;;霜監(jiān)測模型的探討——以湖南省為例[J];中國農(nóng)學通報;2013年29期

3 張學藝;李劍萍;秦其明;韓穎娟;張曉煜;王連喜;官景得;;幾種干旱監(jiān)測模型在寧夏的對比應用[J];農(nóng)業(yè)工程學報;2009年08期

4 李紅梅;馬玉壽;;基于EOS/MODIS的青海草原春季干旱監(jiān)測模型[J];草業(yè)科學;2008年11期

5 孫麗;王飛;吳全;;干旱遙感監(jiān)測模型在中國冬小麥區(qū)的應用[J];農(nóng)業(yè)工程學報;2010年01期

6 王祥峰;蒙繼華;;基于HJ-1衛(wèi)星的農(nóng)田土壤有機質(zhì)含量監(jiān)測[J];農(nóng)業(yè)工程學報;2014年08期

7 匡昭敏;鐘仕全;黃永t，

本文編號：1365081