【摘要】：低空風(fēng)切變已經(jīng)成為嚴(yán)重影響飛機(jī)起飛和進(jìn)場(chǎng)著陸階段的一個(gè)危險(xiǎn)因素,而且不同風(fēng)切變具有獨(dú)特的風(fēng)場(chǎng)特征,對(duì)飛機(jī)的飛行影響也有很大不同,需要飛行員針對(duì)不同類型的風(fēng)切變做出最正確的操作,所以對(duì)低空風(fēng)切變的探測(cè)以及對(duì)其類型的識(shí)別就十分重要。文中利用仿真的二維激光雷達(dá)風(fēng)切變圖像,對(duì)微下?lián)舯┝�、低空急流、海陸風(fēng)、側(cè)風(fēng)四種常見的低空風(fēng)切變進(jìn)行了類型的識(shí)別研究。具體工作如下:第一,針對(duì)激光雷達(dá)實(shí)測(cè)的風(fēng)切變資料的缺乏,文中采用基于Fluent的數(shù)值模擬方法對(duì)四種低空風(fēng)切變進(jìn)行了仿真,可以得到理想條件四種風(fēng)切變的基本特征;并參考香港機(jī)場(chǎng)激光雷達(dá)的探測(cè)方式,同時(shí)采用直接對(duì)仿真風(fēng)場(chǎng)旋轉(zhuǎn)的方式得到了不同風(fēng)向下,激光雷達(dá)測(cè)得的二維風(fēng)切變圖像。第二,針對(duì)不同風(fēng)向下,風(fēng)速在激光雷達(dá)徑向投影的變化以及每種風(fēng)切變所固有的強(qiáng)弱切變區(qū)域,提出了一種組合LBP局部紋理特征和灰度-梯度共生矩陣全局紋理特征的識(shí)別方法。LBP局部紋理特征對(duì)風(fēng)速徑向投影的變化不敏感,灰度-梯度共生矩陣特征全局紋理特征,代表風(fēng)切變整體的切變強(qiáng)弱關(guān)系。再通過典型相關(guān)分析對(duì)兩種特征進(jìn)行融合得到二者的組合紋理特征,最后應(yīng)用支持向量機(jī)對(duì)組合紋理進(jìn)行了分類識(shí)別。最終實(shí)驗(yàn)驗(yàn)證了該方法較單一紋理的識(shí)別具有明顯的提高。第三,針對(duì)標(biāo)準(zhǔn)支持向量機(jī)在激光雷達(dá)風(fēng)切變識(shí)別中不提供后驗(yàn)概率這一問題,提出一種基于有約束FCM的概率支持向量機(jī)建模方法,與傳統(tǒng)方法相比,該方法的性能有一定的提高。
文內(nèi)圖片：
圖片說明：微下?lián)舯┝?br/>[Abstract]:Low altitude wind shear has become a dangerous factor that seriously affects the take-off and approach landing stage of aircraft, and different wind shear has unique wind field characteristics, and the impact on aircraft flight is also very different. It is necessary for pilots to make the most correct operation according to different types of wind shear, so it is very important to detect low altitude wind shear and identify its types. In this paper, the simulated two-dimensional lidar wind shear images are used to identify four common low-altitude wind shear types: micro-downburst, low-level jet, sea-land wind and side wind. The specific work is as follows: first, in view of the lack of measured wind shear data of lidar, four kinds of low altitude wind shear can be simulated by using the numerical simulation method based on Fluent, and the basic characteristics of four kinds of wind shear can be obtained by referring to the detection mode of Hong Kong airport lidar, and the two-dimensional wind shear images measured by lidar under different wind directions are obtained by rotating the simulated wind field directly. Secondly, aiming at the variation of wind speed in lidar radial projection under different wind directions and the inherent strong and weak shear regions of each wind shear, a recognition method combining local texture features and global texture features of gray-gradient symbiosis matrix is proposed. LBP local texture features are insensitive to the change of radial projection of wind speed, and the global texture features of gray-gradient symbiosis matrix are global. It represents the relationship between shear strength and strength of wind shear as a whole. Then the combined texture features are obtained by the fusion of the two features by canonical correlation analysis. Finally, the combined texture is classified and recognized by support vector machine (SVM). Finally, the experimental results show that this method is obviously better than the single texture recognition. Thirdly, in order to solve the problem that standard support vector machine does not provide posterior probability in lidar wind shear recognition, a probabilistic support vector machine modeling method based on constrained FCM is proposed. Compared with the traditional method, the performance of this method is improved to a certain extent.
【學(xué)位授予單位】：中國民航大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：V321.225;TN958.98

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 蔣立輝;高志光;熊興隆;姚彬;莊子波;;基于激光雷達(dá)圖像處理的低空風(fēng)切變類型識(shí)別研究[J];紅外與激光工程;2012年12期

2 陶嵐;戴建華;;下?lián)舯┝髯詣?dòng)識(shí)別算法研究[J];高原氣象;2011年03期

3 王楠;劉黎平;徐寶祥;張建云;;利用多普勒雷達(dá)資料識(shí)別低空風(fēng)切變和輻合線方法研究[J];應(yīng)用氣象學(xué)報(bào);2007年03期

4 鐘志慶;孫東松;王邦新;夏海云;董晶晶;周小林;周軍;;基于Fabry-Perot標(biāo)準(zhǔn)具的多普勒測(cè)風(fēng)激光雷達(dá)[J];紅外與激光工程;2006年06期

5 張翔;肖小玲;徐光yP;;基于最大熵估計(jì)的支持向量機(jī)概率建模[J];控制與決策;2006年07期

6 楊義彬;;激光雷達(dá)技術(shù)的發(fā)展及其在大氣環(huán)境監(jiān)測(cè)中的應(yīng)用[J];成都信息工程學(xué)院學(xué)報(bào);2005年06期

7 杜樹新,吳鐵軍;模式識(shí)別中的支持向量機(jī)方法[J];浙江大學(xué)學(xué)報(bào)(工學(xué)版);2003年05期

8 洪繼光;;灰度-梯度共生矩陣紋理分析方法[J];自動(dòng)化學(xué)報(bào);1984年01期

，

本文編號(hào)：2510585