| CPC G01S 13/9021 (2019.05) | 3 Claims |
|
1. A method for evaluating anti-skid performance of airport pavement based on spaceborne synthetic aperture radar, synthetic aperture radar, abbreviated as SAR, the method comprises following steps:
S1. basing on the synthetic aperture radar images and digital elevation model, obtaining a multi-look intensity raster image with echo intensity values corresponding to scatter points;
obtaining single look complex images and SAR precise orbit data through a SAR; extracting a common area and designating the research scopes; inputting SAR precision orbit data, and combining it with the digital elevation model for image registration to obtain a ground echo intensity scatter plot containing geographic coding and multiple SAR acquisition times;
calculating echo intensity related indicators for each scatter point one by one, the echo intensity related indicators comprising amplitude mean and amplitude deviation; and selecting scatter points that meet screening conditions of amplitude mean and amplitude deviation; after filtering, finally, the ground echo intensity scatter plot is rasterized into a multi-look intensity raster image, the multi-look intensity raster image is presented and used for subsequent processing;
S2. selecting research sites, and obtaining type, material, and process of research scopes corresponding to the research sites, performing laser scanning on pavements to obtain surface texture parameters related to a texture depth and measured friction coefficients;
S3. establishing and outputting a SAR image echo intensity feedforward neural network model;
setting input indicators and output indicators;
Input indicators:
meteorological conditions at the time of SAR acquisition;
types of pavements, pavement material data, pavement treatment process parameters, surface texture parameters and measured friction coefficients;
output indicator: echo intensity values corresponding to the research sites in the multi-look intensity raster image,
data of the input indicators and the output indicators is divided into a training set and a test set; the training set is machine-learned to obtain the SAR image echo intensity feedforward neural network model;
S4. basing on the feedforward neural network model, predict echo intensity values are obtained and compared numerically with echo intensity values of SAR images to be predicted, in order to determine a friction coefficient best fit value; basing on the friction coefficient best fit value, the anti-skid performance of airport pavement is evaluated;
The friction coefficient best fit value is determined by the following steps:
obtaining echo intensity values of the SAR images to be predicted and corresponding meteorological conditions for SAR acquisition; setting assumed friction coefficient values;
inputting the echo intensity values of the SAR images to be predicted, the corresponding meteorological conditions, and the assumed friction coefficient values into the SAR image echo intensity feedforward neural network model to obtain predicted echo intensity values;
calculating differences between the predicted echo intensity values and echo intensity values of the SAR images to be predicted; an assumed friction coefficient value corresponding to an predicted echo intensity value with a smallest difference is the friction coefficient best fit value of the SAR images to be predicted.
|