【摘要】：氣候資源是與農(nóng)業(yè)生產(chǎn)最為密切的資源，并在很大程度上影響主要農(nóng)作物的生長，直接影響了廣大農(nóng)民的收入和國家的糧食安全。利用地理信息系統(tǒng)實現(xiàn)小網(wǎng)格點的氣候資源推算和分析，通過詳細(xì)描述氣候資源在不同地形地貌下的空間分布特征，為區(qū)域性的特色農(nóng)業(yè)發(fā)展、農(nóng)業(yè)生產(chǎn)合理布局以及未來農(nóng)業(yè)發(fā)展規(guī)劃提供可靠的科學(xué)依據(jù)。 在精細(xì)化農(nóng)業(yè)氣候區(qū)劃中，需要對大量的歷史數(shù)據(jù)、使用多種插值算法進(jìn)行小網(wǎng)格插值計算，對于海量數(shù)據(jù)的處理及插值算法的復(fù)雜性，需要大量的計算時間，導(dǎo)致系統(tǒng)的反應(yīng)速度很慢，系統(tǒng)整體性能下降。 克里金(Kriging)插值法，是局部插值法,是一種最優(yōu)內(nèi)插法。Kriging插值是建立在變異函數(shù)空間分析基礎(chǔ)上，，對有限區(qū)域內(nèi)的區(qū)域化變量取值進(jìn)行無偏最優(yōu)估計的一種方法。與其它插值方法相比，不僅考慮了待插點與鄰近己知點的空間相關(guān)性，而且能夠給出估計誤差。用該方法對湖南小網(wǎng)格點氣溫數(shù)據(jù)進(jìn)行海量插值，由于串行Kriging算法比較復(fù)雜，且循環(huán)較多導(dǎo)致計算量大，且小網(wǎng)格點個數(shù)達(dá)到107級會導(dǎo)致計算量較大，計算時間較長。在基于傳統(tǒng)的單處理器模式，當(dāng)處理大規(guī)模海量數(shù)據(jù)時，由于計算時間太長，無法滿足實時分析的需求。雖然在高性能計算機或者分布式計算機中得以解決，但是在PC中無法滿足實時計算的需求。 多核并行計算技術(shù)是計算技術(shù)發(fā)展的重要方向之一。多核計算是使用并行處理技術(shù)進(jìn)行編程，開發(fā)并行性、同時執(zhí)行多個任務(wù)，為合理地提高多核處理器性能提供一個理想的平臺，也是提升系統(tǒng)性能的關(guān)鍵技術(shù)之一，使目前計算機的處理水平有一個質(zhì)的飛躍。因此，使用多線程編程技術(shù)和基于共享存儲的OpenMP編程模型，對串行Kriging算法進(jìn)行改進(jìn)，不僅改善了該算法效率，而且大量實驗數(shù)據(jù)也表明改進(jìn)的Kriging算法的性價比很高，滿足了對湖南地區(qū)小網(wǎng)格實時插值的需求。 采用了多核并行計算技術(shù)所構(gòu)建的信息系統(tǒng)，通過多線程編程技術(shù)，充分利用硬件資源，使得并行計算的硬件資源在信息系統(tǒng)開發(fā)過程中真正發(fā)揮了作用，大大提高了系統(tǒng)的反應(yīng)速度和整體性能表現(xiàn)。
[Abstract]:Climate resources are the most closely related to agricultural production, and to a large extent affect the growth of major crops, directly affect the income of farmers and national food security. The geographic information system (GIS) is used to calculate and analyze the climate resources of small grid points. By describing the spatial distribution characteristics of climate resources in different landforms and landforms in detail, it is a regional characteristic agriculture development. Reasonable distribution of agricultural production and future agricultural development plan provide reliable scientific basis. In fine agricultural climate regionalization, a large number of historical data are needed, and a variety of interpolation algorithms are used to carry out small grid interpolation calculation. For the processing of massive data and the complexity of interpolation algorithm, it takes a lot of computing time. As a result, the response speed of the system is very slow and the overall performance of the system is reduced. Kriging (Kriging) interpolation method is a local interpolation method. Kriging interpolation is an unbiased optimal method for estimating the values of regionalized variables in a finite region based on the spatial analysis of variogram. Compared with other interpolation methods, not only the spatial correlation between the points to be inserted and the nearest known points is considered, but also the estimation error is given. This method is used to interpolate the temperature data of small grid points in Hunan province. Because of the complexity of serial Kriging algorithm and the large amount of calculation due to more cycles, the number of small grid points reaching 107-level will lead to a large amount of calculation and a longer calculation time. Based on the traditional single-processor mode, when processing massive data, the computation time is too long to meet the needs of real-time analysis. Although it can be solved in high performance computer or distributed computer, it can not meet the demand of real-time computing in PC. Multi-core parallel computing technology is one of the important directions in the development of computing technology. Multi-core computing is to use parallel processing technology to program, develop parallelism and perform multiple tasks at the same time. It provides an ideal platform for reasonably improving the performance of multi-core processors, and is also one of the key technologies to improve system performance. Make the current computer processing level has a qualitative leap. Therefore, using multi-thread programming technology and OpenMP programming model based on shared storage, the serial Kriging algorithm is improved, which not only improves the efficiency of the algorithm, but also shows that the improved Kriging algorithm has a high performance-price ratio. It meets the demand of small grid interpolation in Hunan. The information system based on multi-core parallel computing technology is adopted. Through multi-thread programming technology, the hardware resources of parallel computing are fully utilized, which makes the hardware resources of parallel computing really play a role in the development of information system. The reaction speed and the overall performance of the system are greatly improved.
【學(xué)位授予單位】：湖南大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2012
【分類號】：TP338.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 曾懷恩;黃聲享;;基于Kriging方法的空間數(shù)據(jù)插值研究[J];測繪工程;2007年05期

2 馬軒龍;李春娥;陳全功;;基于GIS的氣象要素空間插值方法研究[J];草業(yè)科學(xué);2008年11期

3 傅抱璞;地形和海拔高度對降水的影響[J];地理學(xué)報;1992年04期

4 封志明,楊艷昭,丁曉強,林忠輝;氣象要素空間插值方法優(yōu)化[J];地理研究;2004年03期

5 劉宇;陳泮勤;張穩(wěn);胡非;;一種地面氣溫的空間插值方法及其誤差分析[J];大氣科學(xué);2006年01期

6 李新,程國棟,盧玲;空間內(nèi)插方法比較[J];地球科學(xué)進(jìn)展;2000年03期

7 彭楠峰;;距離反比插值算法與Kriging插值算法的比較[J];大眾科技;2008年05期

8 殷劍敏;辜曉青;林春;;寒露風(fēng)災(zāi)害評估的空間分析模型研究[J];氣象與減災(zāi)研究;2006年03期

9 蘇姝,林愛文,劉慶華;普通Kriging法在空間內(nèi)插中的運用[J];江南大學(xué)學(xué)報;2004年01期

10 陳鵬;王乘;任波;;并行Kriging地層電性參數(shù)分布估計[J];計算機工程與應(yīng)用;2008年15期

相關(guān)碩士學(xué)位論文 前1條

1 程志;多核并行計算在視頻服務(wù)中的研究及應(yīng)用[D];山東大學(xué);2008年

本文編號：2177406