Beidou Zhang

A Novel Algorithm of Image Enhancement Based on Pulse Coupled Neural Network Time Matrix and Rough Set

This paper describes a novel algorithm of image enhancement based on pulse coupled neural network (PCNN) time matrix and rough set indiscernibility relation. Firstly, detect image noise using PCNN time matrix, and then partition the original image into three sub-images according to intensity attribute and noise attribute. Secondly, denoise using filtering methods based upon PCNN. Lastly, complete sub-images and enhance each of them by different methods. For the gray images with many object details in dark regions and badly corrupted by impulse noise, the computer simulations show excellent enhancement effect. Namely, noise can be reduced efficiently, object details can be enhanced better and the image would become clear after it is processed by this algorithm. Moreover, the effect of this algorithm is better than that of traditional image enhancement algorithm.

Turbulence regimes and the validity of similarity theory in the stable boundary layer over complex terrain of the Loess Plateau, China

To gain an insight into the characteristics of turbulence in a stable boundary layer over the complex terrain of the Loess Plateau, data from the Semi-Arid Climate and Environment Observatory of Lanzhou University are analyzed. We propose a method to identify and efficiently isolate nonstationary motions from turbulence series, and then we examine the characteristics of nonstationary motions (nonstationary motions refer to gusty events on a greater scale than local shear-generated turbulence). The occurrence frequency of nonstationary motions is found to depend on the mean flow, being more frequent in weak wind conditions and vanishing when the wind speed, U, is greater than 3.0 ms(-1). When U exceeds the threshold value of 1.0 ms(-1) for the gradient Richardson number Ri 0.3, local shear-generated turbulence on timescales of less than 4min depends systematically on U with an average rate of 0.05 U. However, for the weak wind condition, neither the mean wind speed nor the stability is an important factor for local turbulence. Then turbulence is categorized into three regimes based on the behaviors of nonstationary motions and local turbulence. Regime 1 considers stationary turbulence with a wind speed greater than 3.0 ms(-1), and the Monin-Obukhov similarity theory (MOST) can be used to calculate the turbulence momentum flux. Regime 2 examines intermittent turbulence where the MOST is competent to evaluate the local turbulence momentum flux but not nonstationary motions. Regime 3 involves wind speed that is less than the threshold value, where nonstationary motions are dominant, local turbulence is independent of the mean flow, and where the MOST may well be invalid.

Lidar measurement of planetary boundary layer height and comparison with microwave profiling radiometer observation

The paper is on the determination of the height of the planetary boundary layer (BLH) by means of lidar measurements and application of the continuous wavelet transform method. The retrieved heights are compared to results from numerical models based on the parcel method. The latter allows to determine the entrainment zone; the required information concerning the surface heat flux and the temperature profile are provided from a microwave radiometer and sonic anemometer. The authors retrieve a set of BLHs for Lanzhou and Yuzhong.

A Novel Algorithm of Image Gaussian Noise Filtering based on PCNN Time Matrix

The problem of image Gaussian noise filtering in the framework of Pulse Coupled Neural Network (PCNN) time matrix is addressed. The time matrix, generated by PCNN, contains useful information related to spatial structure of the image under processing. It is a mapping from image spatial information to time sequence. Through time matrix, Gaussian noisy pixels can be detected and then processed by using five methods respectively. Computer simulations show that Gaussian noise can be reduced efficiently, and visual effect of restored images by using the proposed algorithm is much better than those by using traditional noise reduction methods, such as Median Filter, Mean Filter and even Wiener Filter. The proposed algorithm presents higher Peak Signal-to-Noise Ratio, better capability to reduce noise and better protection to edges and details of images. It is a novel Gaussian noise filtering method, which is comparable to Wiener Filter.

An overview of passive and active dust detection methods using satellite measurements

In this paper, the methods to detect dust based on passive and active measurements from satellites have been summarized. These include the visible and infrared (VIR) method, thermal infrared (TIR) method, microwave polarized index (MPI) method, active lidar-based method, and combined lidar and infrared measurement (CLIM) method. The VIR method can identify dust during daytime. Using measurements at wavelengths of 8.5, 11.0, and 12.0 µm, the TIR method can distinguish dust from other types of aerosols and cloud, and identify the occurrence of dust over bright surfaces and during night. Since neither the VIR nor the TIR method can penetrate ice clouds, they cannot detect dust beneath ice clouds. The MPI method, however, can identify about 85% of the dust beneath ice clouds. Meanwhile, the active lidar-based method, which uses the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) data and five-dimensional probability distribution functions, can provide very high-resolution vertical profiles of dust aerosols. Nonetheless, as the signals from dense dust and thin clouds are similar in the CALIOP measurements, the lidar-based method may fail to distinguish between them, especially over dust source regions. To address this issue, the CLIM method was developed, which takes the advantages of both TIR measurements (to discriminate between ice cloud and dense dust layers) and lidar measurements (to detect thin dust and water cloud layers). The results obtained by using the new CLIM method show that the ratio of dust misclassification has been significantly reduced. Finally, a concept module for an integrated multi-satellites dust detection system was proposed to overcome some of the weaknesses inherent in the single-sensor dust detection.

Compact printed dipole antenna with folding element for 2.4 Ghz WLAN communications

A novel printed dipole antenna suitable for 2.4 GHz WLAN application is presented and investigated in this paper. A folding structure in the end of dipole antenna is designed so that the antenna length has a reduction of 46.43 % compared to traditional printed dipole antenna. The effects of some important parameters on the return loss of the proposed antenna have been investigated. In order to operate as WLAN device, the printed dipole antenna should be characterized by a low profile and cover the operation bandwidth of the 2.400-2.484 GHz. The radiation pattern is similar to those conventional dipole antennas and the measured gain variation is less than 0.3dBi.

Evaluation of retrieval methods of daytime convective boundary layer height based on lidar data

The atmospheric boundary layer height is a basic parameter in describing the structure of the lower atmosphere. Because of their high temporal resolution, ground-based lidar data are widely used to determine the daytime convective boundary layer height (CBLH), but the currently available retrieval methods have their advantages and drawbacks. In this paper, four methods of retrieving the CBLH (i.e., the gradient method, the idealized backscatter method, and two forms of the wavelet covariance transform method) from lidar normalized relative backscatter are evaluated, using two artificial cases (an idealized profile and a case similar to real profile), to test their stability and accuracy. The results show that the gradient method is suitable for high signal-to-noise ratio conditions. The idealized backscatter method is less sensitive to the first estimate of the CBLH; however, it is computationally expensive. The results obtained from the two forms of the wavelet covariance transform method are influenced by the selection of the initial input value of the wavelet amplitude. Further sensitivity analysis using real profiles under different orders of magnitude of background counts show that when different initial input values are set, the idealized backscatter method always obtains consistent CBLH. For two wavelet methods, the different CBLH are always obtained with the increase in the wavelet amplitude when noise is significant. Finally, the CBLHs as measured by three lidar-based methods are evaluated by as measured from L-band soundings. The boundary layer heights from two instruments coincide with m in most situations.

Pathologic region detection algorithm for prostate ultrasonic image based on PCNN

It is quite important and difficult for doctors to detect pathologic regions of prostate ultrasonic images. An automated region detection algorithm is proposed to solve this problem, especially for ultrasonic images containing all kinds of noise and speckle. First, all the pixels of an ultrasonic image are fired by Pulse Coupled Neural Network (PCNN). Then after being processed by morphological closing, binary reversing and region labeling, the seeds are detected automatically using PCNN, by which the region of interest (ROI) of the ultrasonic image is detected by Region Growing. In the end, we code the ROI by pseudo-color. Detected pathologic regions can be used for further clinical inspection and quantitative analysis of ultrasonic images.

Automated detection of cloud and aerosol features with SACOL micro-pulse lidar in northwest China

The detection of cloud and aerosols using a modified retrieval algorithm solely for a ground-based micropulse lidar (MPL) is presented, based on one-year data at the Semi-Arid Climate Observatory and Laboratory (SACOL) site (35.57°N, 104.08°E, 1965.8 m), northwest of China, from March 2011 to February 2012. The work not only identifies atmosphere particle layers by means of the range-dependent thresholds based on elastic scattering ratio and depolarization ratio, but also discriminates the detected layers by combining empirical thresholds of the atmospheres thermodynamics states and scattering properties and continuous wavelet transform (CWT) analyses. Two cases were first presented in detail that demonstrated that the modified algorithm can capture atmosphere layers well. The cloud macro-physical properties including cloud base height (CBH), cloud geometrical thickness (CGT), and cloud fraction (CF) were then analyzed in terms of their monthly and seasonal variations. It is shown that the maximum/minimum CBHs were found in summer (4.66 ± 1.95km)/autumn (3.34 ± 1.84km). The CGT in winter (1.05 ± 0.43km) is slightly greater than in summer (0.99 ± 0.44km). CF varies significantly throughout year, with the maximum value in autumn (0.68), and a minimum (0.58) in winter, which is dominated by single-layered clouds (81%). The vertical distribution of CF shows a bimodal distribution, with a lower peak between 1 and 4km and a higher one between 6and 9km. The seasonal and vertical variations in CF are important for the local radiative energy budget.

Enhanced Bottom-of-the-Atmosphere Cooling and Atmosphere Heating Efficiency by Mixed-Type Aerosols: A Classification Based on Aerosol Nonsphericity

AbstractThe current understanding of the climate effects of mixed-type aerosols is an open question. The optical and radiative properties of the anthropogenic, mixed-type, and dust aerosols were studied using simultaneous observations of a sun photometer and a depolarization lidar over the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL), northwestern China. The aerosol radiative effect was calculated using the Santa Barbara DISORT Atmospheric Radiative Transfer (SBDART) model and was in good agreement with the Aerosol Robotic Network (AERONET) product. The anthropogenic, mixed-type, and dust aerosols were identified mainly based on the lidar-measured depolarization ratio, which was supported by the airmass back trajectories. The mixed-type aerosols exhibit lower (higher) extinctions below (above) 1.5 km above the ground, indicating anthropogenic pollution from the atmospheric boundary layer and dust aerosols above. The dust aerosols exhibit the highest absolute radiative effect...