基于激光雷达的风切变识别的研究
发布时间：2019-07-05 14:18
【摘要】：低空风切变已经成为严重影响飞机起飞和进场着陆阶段的一个危险因素,而且不同风切变具有独特的风场特征,对飞机的飞行影响也有很大不同,需要飞行员针对不同类型的风切变做出最正确的操作,所以对低空风切变的探测以及对其类型的识别就十分重要。文中利用仿真的二维激光雷达风切变图像,对微下击暴流、低空急流、海陆风、侧风四种常见的低空风切变进行了类型的识别研究。具体工作如下:第一,针对激光雷达实测的风切变资料的缺乏,文中采用基于Fluent的数值模拟方法对四种低空风切变进行了仿真,可以得到理想条件四种风切变的基本特征;并参考香港机场激光雷达的探测方式,同时采用直接对仿真风场旋转的方式得到了不同风向下,激光雷达测得的二维风切变图像。第二,针对不同风向下,风速在激光雷达径向投影的变化以及每种风切变所固有的强弱切变区域,提出了一种组合LBP局部纹理特征和灰度-梯度共生矩阵全局纹理特征的识别方法。LBP局部纹理特征对风速径向投影的变化不敏感,灰度-梯度共生矩阵特征全局纹理特征,代表风切变整体的切变强弱关系。再通过典型相关分析对两种特征进行融合得到二者的组合纹理特征,最后应用支持向量机对组合纹理进行了分类识别。最终实验验证了该方法较单一纹理的识别具有明显的提高。第三,针对标准支持向量机在激光雷达风切变识别中不提供后验概率这一问题,提出一种基于有约束FCM的概率支持向量机建模方法,与传统方法相比,该方法的性能有一定的提高。
文内图片：
图片说明：微下击暴流
[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.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：V321.225;TN958.98
本文编号：2510585
文内图片：
图片说明：微下击暴流
[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.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：V321.225;TN958.98
【参考文献】
相关期刊论文 前8条
1 蒋立辉;高志光;熊兴隆;姚彬;庄子波;;基于激光雷达图像处理的低空风切变类型识别研究[J];红外与激光工程;2012年12期
2 陶岚;戴建华;;下击暴流自动识别算法研究[J];高原气象;2011年03期
3 王楠;刘黎平;徐宝祥;张建云;;利用多普勒雷达资料识别低空风切变和辐合线方法研究[J];应用气象学报;2007年03期
4 钟志庆;孙东松;王邦新;夏海云;董晶晶;周小林;周军;;基于Fabry-Perot标准具的多普勒测风激光雷达[J];红外与激光工程;2006年06期
5 张翔;肖小玲;徐光yP;;基于最大熵估计的支持向量机概率建模[J];控制与决策;2006年07期
6 杨义彬;;激光雷达技术的发展及其在大气环境监测中的应用[J];成都信息工程学院学报;2005年06期
7 杜树新,吴铁军;模式识别中的支持向量机方法[J];浙江大学学报(工学版);2003年05期
8 洪继光;;灰度-梯度共生矩阵纹理分析方法[J];自动化学报;1984年01期
，本文编号：2510585
本文链接：https://www.wllwen.com/kejilunwen/wltx/2510585.html
