System and method for polarimetric wavelet fractal detection and imaging
A system and method for detection of a target object/material includes identifying a polarimetric signal for a plurality of aspect angles. One/two-dimensional Mueller matrix image or one/two-dimensional Stokes vector image can be processed using power spectral analysis, wavelet and fractal analysis for further image, having increased discrimination with reduced false-ratio. In addition, each of the angular polarization states due to their association with a particular aspect angle are then cross-correlated to generate a two-dimensional image that relates the level of correlation with the aspect angle. Finally, the output information, including statistical parameters are fed to the input of a neural-fuzzy network for further optimization and image enhancement.
1. A method for imaging a target object illuminated by a light source at a plurality of aspect angles, the method comprising: generating an angular Mueller Matrix for each one of the aspect angles measured using the light source, said Mueller Matrix having a plurality of matrix elements each representing a polarimetric state of the target object at one of the aspect angles, said matrix elements being expressed as an analog signal; generating a Stokes vector for each one of the plurality of aspect angles, wherein said Stokes vector includes a plurality of vector elements; representing each one of said plurality of matrix elements and each one of said vector elements as a wavelet element; identifying a low frequency component and a high frequency component of each said wavelet element; estimating a power spectral density from said high and low frequency components; estimating a fractal dimension of said wavelet elements of each said Mueller Matrix and each said Stokes vector; generating a cross-correlation based on a polarimetric state of each said Mueller Matrix; and displaying a two-dimensional image of the target object based on said power spectral density, said wavelet elements, said cross-correlation, and said fractal dimension.
2. The method of claim 1 , wherein the light source is active or passive.
3. The method of claim 1 , wherein the light source is under transmission or backscattered geometry.
4. The method of claim 1 , wherein said representing step is performed by a continuous wavelet transform.
5. The method of claim 4 , wherein said continuous wavelet transform comprises a Daubechies wavelet transform.
6. The method of claim 1 , wherein said fractal dimension is estimated by a method selected from the group consisting of: a box counting method, a multi-resolution box-counting method, a Katz method, a Sevcik method, a Higuhi method, a regularization method, or a maximum entropy method.
7. A method for imaging a target object illuminated by a light source at a plurality of aspect angles, the method comprising: generating an angular Mueller Matrix for each one of the aspect angles measured using the light source, said Mueller Matrix having a plurality of matrix elements each representing a polarimetric state of the target object at one of the aspect angles, said matrix elements being expressed as an analog signal; representing said matrix elements as respective wavelet elements; identifying a low frequency component and a high frequency component of said wavelet elements; estimating a power spectral density from said high and low frequency components; estimating a fractal dimension of said wavelet elements of each said Mueller Matrix; generating a cross-correlation based on a polarimetric state of each said Mueller Matrix; and displaying an image of the target object based on said power spectral density, said wavelet elements, said cross-correlation, and said fractal dimension.
8. The method of claim 7 , wherein the light source is active or passive.
9. The method of claim 7 , wherein the light source is under transmission or backscattered geometry.
10. The method of claim 7 , further comprising:
optimizing said displaying step by processing said power spectral density and said fractal dimension through a neural-fuzzy network.
11. The method of claim 10 , wherein said neural-fuzzy network comprises a two-layer feed-forward network.
12. The method of claim 7 , wherein said continuous wavelet transform comprises a Daubechies wavelet transform.
13. The method of claim 7 , wherein said fractal dimension is estimated by a method selected from the group consisting of: a box counting method, a multi-resolution box-counting method, a Katz method, a Sevcik method, a Higuhi method, a regularization method, or a maximum entropy method.