-
通过检测小波变换模极大值来检测信号奇异点的方法最早是由MALLAT等人提出[12],后人对此加以引申并应用到了电力系统故障检测[13]、心电图异常信号检测[14-15]、桥梁裂缝检测[16]。在风切变的检测中,风切变可以看成是径向速度信号的突变信号。信号发生的突变时刻被认为是信号的奇异点,而小波变换模极大值通常正好对应着信号的突变点,因此,基于小波变换模极大值检测方法可用于低空风切变预警,通过检测速度径向数据的小波变换模极大值确定信号的突变位置,即检测飞机下滑道上风切变发生的时刻和位置。
MALLAT系统地论述了如何利用小波变换的局部化特性检测信号奇异点位置[12],并对小波变换的定义采用了具有滤波意义的卷积形式[17]。
设ψ(x)为一个小波母函数,满足容许性条件,尺度为a的小波母函数为$ \psi_a(x)=\frac{1}{a} \psi\left(\frac{x}{a}\right)$,其小波变换为f(x)和ψa(x)的卷积如(1)式所示,其在物理意义上表现为脉冲响应为ψa(x)的滤波器对信号f(x)的滤波:
$ \begin{gathered} W_\psi(a, x)=f(x) * \psi_a(x)= \\ \frac{1}{a} \int_{-\infty}^{+\infty} f(x) \psi\left(\frac{x-\tau}{a}\right) \mathrm{d} \tau \end{gathered} $
(1) 式中,τ为时间。
设具有低通平滑作用的滤波函数θ(x)满足:$ \int_{-\infty}^{+\infty} \theta(x) \mathrm{d} x=1$和$ \lim\limits _{x \mapsto \rightarrow \infty} \theta(x)=0$,其1阶导数和2阶导数分别用ζ(x)和η(x)表示:
$ \zeta(x)=\frac{\mathrm{d} \theta(x)}{\mathrm{d} x} $
(2) $ \eta(x)=\frac{\mathrm{d}^2 \theta(x)}{\mathrm{d} x} $
(3) 根据Fourier变换的微分性质,ζ(x),η(x)均满足容许性条件,均可作为小波母函数。对其分别做小波变换,有:
$ \begin{gathered} W_\zeta(a, x)=\frac{1}{a} \int_{-\infty}^{+\infty} f(x) \zeta\left(\frac{x-\tau}{a}\right) \mathrm{d} \tau= \\ a \frac{\mathrm{d}}{\mathrm{d} x}\left[f(x) * \theta_a(x)\right] \end{gathered} $
(4) $ \begin{gathered} W_\eta(a, x)=\frac{1}{a} \int_{-\infty}^{+\infty} f(x) \eta\left(\frac{x-\tau}{a}\right) \mathrm{d} \tau= \\ a^2 \frac{\mathrm{d}^2}{\mathrm{~d} x^2}\left[f(x) * \theta_a(x)\right] \end{gathered} $
(5) 式中,Wζ(a, x)是f(x)通过滤波器θa(x)滤波后的1阶导数。由于θa(x)是一个平滑滤波函数,所以f(x)经过θa(x)滤波后,f(x)的噪声得到了抑制;而1阶导数,即微分运算,反映了f(x)的变化率,当存在突变点时,它的变化率就很大,达到模极大值,所以Wζ(a, x)取极值点的地方就是f(x)的边沿位置。Wη(a, x)是通过滤波器θa(x)滤波后的2阶导数,2阶导数的过零点对应1阶导数的极值点,所以也常用Wη(a, x)的过零点来检测信号的突变点,但是对于受到强噪声污染的信号,Wη(a, x)的过零点很多,由此很难真正的判断信号的边沿。此外,Wη(a, x)的过零点只能给出拐点的位置信息,不能给出变换的快慢,所以本文中用Wζ(a, x)的模极大值来检测径向速度信号的风切变。若对x0的任意邻域内的任意点x有|W(a0, x)|≤|W(a0, x0)|,则称点(a0, x)为小波变换的模极大值。
图 1是本文中提出的新算法的流程图。
基于小波变换模极大值的LiDAR风切变预警算法
Application of LiDAR based on wavelet transform modulus maxima in low-level wind shear alerting
-
摘要: 为了更好地检测低空风切变, 保障飞机的飞行安全, 提出了一种基于小波变换模极大值的激光雷达风切变预警算法。先用小波变换求取重组逆风廓线上的模极大值, 找到"拐点"后再利用风切变判断标准来判断。在进行数值仿真和来自湖北郧西气象站、四川攀枝花机场的现场检测验证后, 确认新算法在准确性及效率方面都有良好的性能, 且脉冲型数据使用biorthogonal系中双数小波检测结果较准确, 而阶跃型和斜坡型数据需使用Daubechies系中Db5小波。结果表明, 鄂西北郧西县和攀枝花保安营机场均有风切变发生, 风切变强度为重度。该算法能检测不同类型的风切变, 不用考虑风切变的尺度, 较好地弥补了现有算法的不足, 为飞机的起降提供技术保障, 对实时检测和预警也有重大的意义。Abstract: For better detection of low-level wind shear, a new algorithm based on the wavelet transform modulus maxima methodI was introduced to predict the occurrence of wind shear along the glide path. Wavelet transform is used to obtain the modulus maximum value on the recombined upwind profile. The "inflection point" was found, and then the windshear judgment standard was used to judge its accuracy. Numerical examples and field detection data from Yunxi Meteorological Station in Hubei Province and Panzhihua Airport in Sichuan Province have well verified the good performance of the method, in terms of both accuracy and efficiency. The result shows that the pulse-type data is more accurate using the even number wavelet in the biorthogonal system; the step-type and ramp-type data is more accurate using the Db5 in the Daubechies system. The results show that the wind shear occurred in Yunxi county and Panzhihua Baoanying Airport, the wind shear intensity were both heavy. This algorithm can detect different types of wind shear without considering the scale of wind shear, which makes up for the shortcomings of existing algorithms, provides technical support for aircraft takeoff and landing, and has great significance for real-time detection and early warning.
-
[1] HUANG T J, TIAN Y H, LI J, et al. Salient region detection and segmentation for general object recognition and image understanding[J]. Science China Information Sciences, 2011, 54(12): 2461-2470. doi: 10.1007/s11432-011-4487-1 [2] ZHANG T, LI Q, ZHENG J F, et al. A study on low-level wind shear caused by microburst using lidar and other data[J]. Laser Technology, 2020, 44(5): 563-569 (in Chinese). [3] INTERNATIONAL CIVIL AVIATION ORGANIZATION. Manual on low-level wind shear[M]. Montreal, Canada: ICAO Headquarters, 2005: 127. [4] FUJITA T T, CARACENA F. An analysis of three weather-related aircraft accidents[J]. Bulletin of American Meteorological Society, 1977, 58(11): 1164-1181. doi: 10.1175/1520-0477(1977)058<1164:AAOTWR>2.0.CO;2 [5] WOODFIELD A A, WOODS J F. Worldwide experience of wind shear during 1981~1982[EB/OL]. (2012-04-24)[2012-12-20]. https://kns-cnki-net-s.nudtproxy.yitlink.com/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFD2012&filename=JGHW201212003&uniplatform=NZKPT&v=iXeUOKtQ33zymOZ6%25mmd2Frl6YHL6SCOvg%25mmd2BrNBO1TjIe15psob5o9AYmVkT3ZgFBgkPza. [6] CHAN P W, HON K K, SHIN D K. Combined use of headwind ramps and gradients based on LiDAR data in the alerting of low-level windshear/turbulence[J]. Meteorologische Zeitschrift, 2011, 20(6): 661-670. doi: 10.1127/0941-2948/2011/0242 [7] SHUN C M, CHAN P W. Application of an infrared Doppler lidar in detection of wind shear[J]. Journal of Applied Meteorology and Climatology, 2008, 25(5): 637-655. [8] CHAN P W. Application of LIDAR-based F-factor in windshear alerting[J]. Meteorologische Zeitschrift, 2012, 21(2): 193-204. doi: 10.1127/0941-2948/2012/0321 [9] HON K K, CHAN P W. Application of LiDAR-derived eddy dissipation rate profiles in low-level wind shear and turbulence alerts at Hong Kong International Airport[J]. Royal Meteorological Society, 2013, 21(1): 75-84. [10] JIANG L H, YAN Y, XIONG X L, et al. Doppler LiDAR alerting algorithm of low-level wind shear based on ramps detection[J]. Infrared and Laser Engineering, 2016, 45(1): 0106001(in Chin-ese). doi: 10.3788/irla201645.0106001 [11] MA Y Zh, CHEN N, XIONG X L. Wind shear warning algorithm based on PCA and phase difference correction [J]. Systems Engineering and Electronics, 2019, 42(1): 52-60 (in Chinese). [12] MALLAT S, HWANG W L. Singularity detection and processing with wavelets[J]. IEEE Transactions on Information Theory, 1992, 38(2): 617-643. doi: 10.1109/18.119727 [13] WANG Y, YIN X, XU W, et al. Fault line selection in cooperation with multi-mode grounding control for the floating nuclear power plant grid[J]. Protection and Control of Modern Power Systems, 2020, 16(5): 1-10. [14] CVETKOVIC D, VBEYLI E D, COSIC I. Wavelet transform feature extraction from human PPG, ECG, and EEG signal responses to ELF PEMF exposures: A pilot study [J]. Digital Signal Processing, 2008, 18(5): 861-874. doi: 10.1016/j.dsp.2007.05.009 [15] CHABCHOU S, MANSOURI S, SALAH R B. Impedance cardiography signal denoising using discrete wavelet transform[J]. Australasian Physical and Engineering Sciences in Medicine, 2016, 39: 655-663. doi: 10.1007/s13246-016-0460-z [16] GENTILE A, MESSINA A. On the continuous wavelet transforms applied to discrete vibrational data for detecting open cracks in da-maged beams[J]. International Journal of Solids and Structures, 2003, 40(2): 295-315. doi: 10.1016/S0020-7683(02)00548-6 [17] MALLAT S, HWANG W L. Singularity detection and processing with wavelets[J]. IEEE Transactions on Information Theory, 1992, 38(2): 617-643. doi: 10.1109/18.119727 [18] FENG L T, ZHOU J, FAN Q, et al. Three-dimensional lidar for wind shera detection and early warning in civil aviation airport[J]. Acta Photonica Sinica, 2019, 48(5): 0512001 (in Chinese). doi: 10.3788/gzxb20194805.0512001