CN105718957A - Polarized SAR image classification method based on nonsubsampled contourlet convolutional neural network - Google Patents

Abstract

The invention discloses a polarized SAR image classification method based on a nonsubsampled contourlet convolutional neural network, and mainly at solving the problems that influence of speckle noises is hard to avoid and the classification precision is low in the prior art. The method comprises the steps that a polarized SAR image to be classified is denoised; Pauli decomposition is carried out on a polarized scattering matrix S obtained by denoising; image characteristics obtained via Pauli decomposition are combined into a characteristic matrix F, and the characteristic matrix F is normalized and recorded as F1; 22*22 blocks surrounding the F1 are taken for each pixel point to obtain a block based characteristic matrix F2; a training data set and a test data set are selected from the F2; the nonsubsampled contourlet convolutional neural network is established to train the training data set; and the trained nonsubsampled contourlet convolutional neural network is used to classify the test data set. The polarized SAR image classification method improves the expression capability and the classification precision of the features of the polarized SAR image, and can be used for target identification.

Description

The Classification of Polarimetric SAR Image method of non-down sampling contourlet convolutional neural networks

Technical field

The invention belongs to technical field of image processing, be specifically related to a kind of Classification of Polarimetric SAR Image method, can be used for target recognition.

Background technology

Polarization SAR is a kind of high-resolution active-mode active microwave remote sensing imaging radar, have round-the-clock, round-the-clock, resolution high, can the advantage such as side-looking imaging, the information that target is more rich can be obtained.The purpose of Classification of Polarimetric SAR Image is to utilize airborne or the acquisition of borne polarization SAR sensor polarization measurement data to determine the classification belonging to each pixel, has research and using value widely in agricultural, forestry, military affairs, geology, hydrology and ocean etc..

The method that the Classification of Polarimetric SAR Image method commonly used now is based on pixel, namely the feature merely with each pixel itself is classified.Although these methods can retain the details of Pixel-level in image preferably, but the impact due to coherent spot, there is error between measured value and the actual value of single pixel, classification chart is difficult to avoid that there is more isolated pixel and pocket, add classification difficulty.

The existing Polarimetric SAR Image target's feature-extraction method based on scattering properties, including Cloude decomposition, Freeman decomposition etc..

1997, Cloude et al. proposed Cloude and decomposes, and H/ α plane is divided, and characterizes the eigenvalue of polarization data by H and α two and each pixel is turned to the classification of respective regions.H/ alpha taxonomy there is a disadvantage that the division in region is excessively dogmatic, when being distributed in two classes or classes of border when of a sort data, classifier performance will be deteriorated, and another weak point is, when coexist in same region several different atural object time, it is impossible to effectively distinguish；

2004, Lee et al. proposed a kind of based on the Freeman feature extracting method decomposed, and the method can keep all kinds of polarization scattering characteristics, but classification results is subject to the impact of Freeman decomposability, and the universality of this algorithm of polarization data of different-waveband is poor.

These feature extracting methods all do not account for the resolution characteristics multiple dimensioned, many of Polarimetric SAR Image, and the Polarimetric SAR Image that background is complicated hardly results in higher nicety of grading.

Summary of the invention

Present invention aims to the problems referred to above, it is proposed to a kind of Classification of Polarimetric SAR Image method based on non-down sampling contourlet convolutional neural networks, to obtain the characteristics of image with resolution characteristic multiple dimensioned, many, promote nicety of grading.

The thinking of the present invention is: based on convolutional neural networks, image block characteristics is processed, and by introducing non-down sampling contourlet transform in the network, is effectively improved the ability to express of Polarimetric SAR Image feature, and its implementation includes as follows:

(1) Polarimetric SAR Image to be sorted is carried out denoising, obtain the filtered polarization scattering matrix S of Polarimetric SAR Image；

(2) filtered polarization scattering matrix S is carried out Pauli decomposition, Pauli is decomposed obtain odd scattering, even scattering, volume scattering value as the characteristics of image of Polarimetric SAR Image；

(3) Pauli is decomposed the eigenmatrix F based on pixel of the polarization SAR image of image characteristic combination obtained, the corresponding 3 dimension Pauli characteristics of decomposition of each pixel, and the element value in F is normalized between [0,1], it is denoted as F1；

(4) each pixel is taken the block of 22 × 22 around F1, obtain block-based eigenmatrix F2, be i.e. the block of corresponding 3 22 × 22 of each pixel；

(5) from block-based eigenmatrix F2, training dataset and test data set are chosen:

(5a) Polarimetric SAR Image atural object being divided into 15 classes, randomly select N number of markd pixel respectively as training sample D1 from each classification, all the other markd pixels take the integer between 300～700 as test sample T1, N；

(5b) with the marginal point of Canny operator extraction Polarimetric SAR Image, training sample D1 adds the marginal point of Canny operator extraction, namely increase the training sample that confidence level is higher, training dataset D after being updated and test data set T；

(6) structure non-down sampling contourlet convolutional neural networks:

(6a) select the 8 layers of convolutional neural networks being made up of input layer → convolutional layer → pond layer → convolutional layer → pond layer → full articulamentum → full articulamentum → softmax grader, and determine the wave filter size of convolutional neural networks and the Feature Mapping figure of each layer；

(6b) replace the level 2 volume lamination in convolutional neural networks with non-down sampling contourlet transform layer, obtain non-down sampling contourlet convolutional neural networks；

(7) with non-down sampling contourlet convolutional neural networks, training dataset is trained；

(8) utilize the non-down sampling contourlet convolutional neural networks trained that test data set is classified, obtain the pixel class of each pixel in Polarimetric SAR Image test data set.

The present invention has the advantage that compared with prior art

1. the present invention extracts image block characteristics in conjunction with pixel space relativity, reduces coherent spot impact, thus improving nicety of grading.

2. due to the fact that employing non-down sampling contourlet convolutional neural networks, and in convolutional neural networks, introduce the characteristics of image that non-down sampling contourlet transform obtains having resolution characteristic multiple dimensioned, many, thus can better approach original image, improve nicety of grading.

Accompanying drawing explanation

Fig. 1 is the flowchart of the present invention；

Fig. 2 is to the pseudocolour picture after image to be classified denoising in the present invention；

The handmarking of image to be classified is schemed in the present invention by Fig. 3；

Fig. 4 is with the present invention classification results figure to image to be classified.

Detailed description of the invention

Below in conjunction with accompanying drawing to the present invention realize step and experiment effect is described in further detail:

With reference to Fig. 1, the present invention to implement step as follows:

Step 1, carries out denoising to Polarimetric SAR Image to be sorted.

Conventional Polarimetric SAR Image denoising method has the filtering of mean filter, medium filtering, local, exquisite polarization LEE filtering etc., and what the present invention adopted is exquisite polarization LEE filter method, specifically comprises the following steps that

(1a) set exquisiteness polarization LEE filtering sliding window, this sliding window be sized to 5 × 5 pixels；

(1b) by sliding window in the pixel of the Polarimetric SAR Image of input, from left to right, roam from top to bottom, when often roaming a step, by sliding window according to pixel space position, from left to right, 9 subwindows it are divided into from top to bottom successively, each subwindow be sized to 3 × 3 pixels, have overlap between subwindow；

(1c) pixel value of each subwindow correspondence position is averaged, obtained average is constituted the average window of 3 × 3 pixels；

(1d) the gradient masterplate of level, vertical, the four direction of 45 degree and 135 degree is chosen, average window is weighted with four masterplates respectively, obtained weighted results is sought absolute value, selects the maximum in all absolute values, using direction corresponding for this maximum as edge direction；

(1e) from 2, the left and right subwindow of 9 subwindow Zhong Qu center window edge directions, respectively all pixel values in these 2 subwindows are averaged, the average of all pixel values of center window it is individually subtracted by 2 averages obtained, using the subwindow corresponding to value little for absolute value in average difference as direction window, wherein, center window refers to the subwindow of the 3 × 3 of 5 × 5 window center；

(1f) according to formula<1>, the weights of exquisite polarization LEE filtering are obtained:

$b=\frac{\mathrm{var}\left(y\right)-p^2{\mathrm{\&sigma;}}_v^2}{(1+{\mathrm{\&sigma;}}_v^2)\mathrm{var}\left(y\right)},---<1>$

Wherein, b represents the weights of exquisite polarization LEE filtering, and var (y) represents the variance yields of polarization SAR general power image pixel in the window of direction, and y represents the pixel of polarization SAR general power image in the window of direction, p represents the average of all pixels of polarization SAR general power image in the window of directionRepresent the variance yields of the Polarimetric SAR Image coherent speckle noise of input；

(1g) according to formula<2>, the polarization coherence matrix T of filtering after-polarization SAR image center pixel is obtained:

T=w+b (z-w),<2>

Wherein, w represents the average of the polarization coherence matrix of Polarimetric SAR Image pixel in the window of direction, and b represents the weights of exquisite LEE filtering, and z represents the polarization coherence matrix of Polarimetric SAR Image center pixel；

(1h) according to formula<3>, horizontal emission and the scattering component S of level reception can be tried to achieve_HH, Vertical Launch and vertical reception scattering component S_VV, horizontal emission and vertical reception scattering component S_HV:

$\left\{\begin{array}{c}{T}_{11}=\frac{1}{2}|S_{HH}+S_{VV}{|}^2\\ T_{22}=\frac{1}{2}|S_{HH}-S_{VV}{|}^2\\ T_{33}=2|S_{HV}{|}^2\end{array}\right.---<3>$

Wherein, T₁₁、T₂₂、T₃₃For element on the diagonal of the coherence matrix T that polarizes.

Pseudocolour picture after image to be classified denoising is as shown in Figure 2.

Step 2, carries out Pauli decomposition to filtered polarization scattering matrix S, Pauli is decomposed obtain odd scattering, even scattering, volume scattering value as the characteristics of image of Polarimetric SAR Image.

(2a) define basic collision matrix, be called Pauli base: { S_a,S_b,S_c,S_d, formula is as follows:

$\left\{\begin{array}{c}{S}_a=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right]\\ S_b=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 0\\ 0& -1\end{array}\right]\\ S_c=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right]\\ S_d=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}0& -i\\ i& 0\end{array}\right]\end{array}\right.---<4>$

Wherein $\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right]$ Represent odd scattering, $\left[\begin{array}{cc}1& 0\\ 0& -1\end{array}\right]$ Represent even scattering, $\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right]$ Represent volume scattering, $\left[\begin{array}{cc}0& -i\\ i& 0\end{array}\right]$ Representing non-existent type of ground objects, therefore d value is 0；

(2b) the Pauli base defined according to formula<4>, obtains the expression formula of polarization scattering matrix S:

$S=\left[\begin{array}{cc}{S}_{HH}& S_{HV}\\ S_{VH}& S_{VV}\end{array}\right]={\mathrm{aS}}_a+{\mathrm{bS}}_b+{\mathrm{cS}}_c+{\mathrm{dS}}_d---<5>$

The wherein value of a correspondence odd scattering, the value of b correspondence even scattering, c represents the value of volume scattering, and d represents the value of scattering composition corresponding to non-existent type of ground objects；

(2c) solve formula<5>, obtain scattering value a, b, c, d, be denoted as vector form as follows:

$K=\&lsqb;\begin{array}{cccc}a& b& c& d\end{array}\&rsqb;=\frac{1}{\sqrt{2}}{\left(\begin{array}{cccc}{S}_{HH}+S_{VV}& S_{HH}-S_{VV}& S_{HV}+S_{VH}& i(S_{VH}-S_{HV})\end{array}\right)}^T---<6>$

When meeting reciprocity condition S_HV=S_VHTime, formula<6>is reduced to:

$K=\&lsqb;\begin{array}{ccc}a& b& c\end{array}\&rsqb;=\frac{1}{\sqrt{2}}\left(\begin{array}{ccc}{S}_{HH}+S_{VV}& S_{HH}-S_{VV}& 2S_{HV}\end{array}\right),---<7>$

The S that formula<3>is tried to achieve_HH、S_VV、S_HVSubstitution formula<7>, tries to achieve polarization characteristic K.

Step 3, decomposes, by Pauli, the image characteristic combination obtained and becomes eigenmatrix F, and it is normalized.

Constructing an eigenmatrix F, matrix size is set as M1 × M2 × 3, by Pauli decompose obtain odd scattering, even scattering, volume scattering value be assigned to eigenmatrix F, wherein M1 is the length of image to be classified, and M2 is the width of image to be classified；

To eigenmatrix F normalization, adopt characteristic line pantography, namely first obtain the maximum max (F) of eigenmatrix F；Again by each element in eigenmatrix F all divided by maximum max (F), obtain normalized eigenmatrix F1.

Step 4, takes the block of 22 × 22 around F1 to each pixel after normalization, obtains block-based eigenmatrix F2, i.e. the block of corresponding 3 22 × 22 of each pixel, and eigenmatrix F2 is sized to 22 × 22 × (M1 × M2) × 3.

Step 5, chooses training dataset and test data set from block-based eigenmatrix F2.

(5a) Polarimetric SAR Image atural object being divided into 15 classes, randomly select N number of markd pixel respectively as training sample D1 from each classification, all the other markd pixels take the integer between 300～700 as test sample T1, N；

(5b) with the marginal point of Canny operator extraction Polarimetric SAR Image, training sample D1 adds the marginal point of Canny operator extraction, namely increase the training sample that confidence level is higher, training dataset D after being updated and test data set T.

Step 6, constructs non-down sampling contourlet convolutional neural networks.

(6a) select the 8 layers of convolutional neural networks being made up of input layer → convolutional layer → pond layer → convolutional layer → pond layer → full articulamentum → full articulamentum → softmax grader, and determine the wave filter size of convolutional neural networks and the Feature Mapping figure of each layer；

(6b) replace the level 2 volume lamination in convolutional neural networks with non-down sampling contourlet transform layer, obtain non-down sampling contourlet convolutional neural networks, obtain following 8 Rotating fields:

Input layer → non-down sampling contourlet transform layer → pond layer → convolutional layer → pond layer → full articulamentum → full articulamentum → softmax grader；

The parameter of each layer is:

1st layer of input layer: output characteristic mapping graph=3；

2nd layer of non-down sampling contourlet transform layer: output characteristic mapping graph=12；

3rd layer of pond layer: down-sampling yardstick=2；

4th layer of convolutional layer: output characteristic mapping graph=20, filter size=4；

5th layer of pond layer: down-sampling yardstick=2；

6th layer of full articulamentum: output characteristic mapping graph=100；

7th layer of full articulamentum: output characteristic mapping graph=64；

8th layer of softmax grader: output characteristic mapping graph=15.

Step 7, is trained training dataset with non-down sampling contourlet convolutional neural networks.

Using the input as non-down sampling contourlet convolutional neural networks of the eigenmatrix of training dataset, network output layer is corresponding prediction class mark, by solving the error between prediction class mark and the correct class mark of handmarking, and error is carried out back propagation, optimize the weights of non-down sampling contourlet convolutional neural networks, error back propagation mode of the present invention is identical with convolutional neural networks, and the correct class mark of handmarking is as shown in Figure 3.

Step 8, utilizes the non-down sampling contourlet convolutional neural networks trained that test data set is classified, obtains the pixel class of each pixel in Polarimetric SAR Image test data set.

The effect of the present invention can be further illustrated by following emulation experiment:

1. simulated conditions:

Emulation experiment adopts the full polarimetric SAR data in the L-band Holland Flevoland area of NASA/JPL lab A IRSAR system, decomposes, based on Pauli, the image obtained and is sized to 750 × 1024 pixels.

Hardware platform is: Intel (R) Xeon (R) CPUE5-2620,2.00GHz*18, inside saves as 64G.

Software platform is: MATLAB_2014a.

2. emulation content and result:

Test under above-mentioned simulated conditions by the inventive method, namely from each classification of polarization SAR data, 700 markd pixels are randomly selected respectively as training sample, all the other markd pixels are as test sample, training dataset accounts for the 6% of total sample number, obtain the classification results such as Fig. 4, as can be seen from Figure 4: except seldom miscounting a point pixel, the region consistency of classification results is better, and profile is very clear.

Reducing training sample more successively, make training dataset account for the 5% of total sample number, 4%, 3%, the test data set precision of the present invention and convolutional neural networks and model training time are contrasted, result is as shown in table 1:

Table 1

As seen from Table 1, the present invention is when training dataset accounts for the 6% of total sample number, 5%, 4%, 3%, and test data set precision is above convolutional neural networks, and the model training time needed is shorter；When training dataset accounts for the 3% of total sample number, the present invention can obtain the nicety of grading of 92%, and the nicety of grading of convolutional neural networks does not restrain.

To sum up, the present invention is effectively increased the ability to express of Polarimetric SAR Image feature by introducing non-down sampling contourlet transform in convolutional neural networks, promotes nicety of grading, and decreases the model training time.When number of samples is less, with the obvious advantage.

Claims (5)

1. a Classification of Polarimetric SAR Image method for non-down sampling contourlet convolutional neural networks, including:

(1) Polarimetric SAR Image to be sorted is carried out denoising, obtain the filtered polarization scattering matrix S of Polarimetric SAR Image；

(2) filtered polarization scattering matrix S is carried out Pauli decomposition, Pauli is decomposed obtain odd scattering, even scattering, volume scattering value as the characteristics of image of Polarimetric SAR Image；

(3) Pauli is decomposed the eigenmatrix F based on pixel of the polarization SAR image of image characteristic combination obtained, the corresponding 3 dimension Pauli characteristics of decomposition of each pixel, and the element value in F is normalized between [0,1], it is denoted as F1；

(4) each pixel is taken the block of 22 × 22 around F1, obtain block-based eigenmatrix F2, be i.e. the block of corresponding 3 22 × 22 of each pixel；

(5) from block-based eigenmatrix F2, training dataset and test data set are chosen:

(5a) Polarimetric SAR Image atural object being divided into 15 classes, randomly select N number of markd pixel respectively as training sample D1 from each classification, all the other markd pixels take the integer between 300～700 as test sample T1, N；

(5b) with the marginal point of Canny operator extraction Polarimetric SAR Image, training sample D1 adds the marginal point of Canny operator extraction, namely increase the training sample that confidence level is higher, training dataset D after being updated and test data set T；

(6) structure non-down sampling contourlet convolutional neural networks:

(6a) select the 8 layers of convolutional neural networks being made up of input layer → convolutional layer → pond layer → convolutional layer → pond layer → full articulamentum → full articulamentum → softmax grader, and determine the wave filter size of convolutional neural networks and the Feature Mapping figure of each layer；

(6b) replace the level 2 volume lamination in convolutional neural networks with non-down sampling contourlet transform layer, obtain non-down sampling contourlet convolutional neural networks；

(7) with non-down sampling contourlet convolutional neural networks, training dataset is trained；

(8) utilize the non-down sampling contourlet convolutional neural networks trained that test data set is classified, obtain the pixel class of each pixel in Polarimetric SAR Image test data set.

2. the Classification of Polarimetric SAR Image method of non-down sampling contourlet convolutional neural networks according to claim 1, wherein carries out denoising to Polarimetric SAR Image to be sorted in step (1), adopt exquisiteness polarization LEE filter method, and its step is as follows:

(1a) set exquisiteness polarization LEE filtering sliding window, this sliding window be sized to 5 × 5 pixels；

(1b) by sliding window in the pixel of the Polarimetric SAR Image of input, from left to right, roam from top to bottom, when often roaming a step, by sliding window according to pixel space position, from left to right, 9 subwindows it are divided into from top to bottom successively, each subwindow be sized to 3 × 3 pixels, have overlap between subwindow；

(1c) pixel value of each subwindow correspondence position is averaged, obtained average is constituted the average window of 3 × 3 pixels；

(1d) the gradient masterplate of level, vertical, the four direction of 45 degree and 135 degree is chosen, average window is weighted with four masterplates respectively, obtained weighted results is sought absolute value, selects the maximum in all absolute values, using direction corresponding for this maximum as edge direction；

(1e) from 2, the left and right subwindow of 9 subwindow Zhong Qu center window edge directions, respectively all pixel values in these 2 subwindows are averaged, the average of all pixel values of center window it is individually subtracted by 2 averages obtained, using the subwindow corresponding to value little for absolute value in average difference as direction window, wherein, center window refers to the subwindow of the 3 × 3 of 5 × 5 window center；

(1f) according to formula<1>, the weights of exquisite polarization LEE filtering are obtained:

$b=\frac{\mathrm{var}\left(y\right)-p^2{\mathrm{\&sigma;}}_v^2}{(1+{\mathrm{\&sigma;}}_v^2)\mathrm{var}\left(y\right)},---<1>$

Wherein, b represents the weights of exquisite polarization LEE filtering, and var (y) represents the variance yields of polarization SAR general power image pixel in the window of direction, and y represents the pixel of polarization SAR general power image in the window of direction, p represents the average of all pixels of polarization SAR general power image in the window of directionRepresent the variance yields of the Polarimetric SAR Image coherent speckle noise of input.

(1g) according to formula<2>, the polarization coherence matrix T of filtering after-polarization SAR image center pixel is obtained:

T=w+b (z-w),<2>

Wherein, w represents the average of the polarization coherence matrix of Polarimetric SAR Image pixel in the window of direction, and b represents the weights of exquisite LEE filtering, and z represents the polarization coherence matrix of Polarimetric SAR Image center pixel.

(1h) according to formula<3>, horizontal emission and the scattering component S of level reception can be tried to achieve_HH, Vertical Launch and vertical reception scattering component S_VV, horizontal emission and vertical reception scattering component S_HV:

$\left\{\begin{array}{c}{T}_{11}=\frac{1}{2}|S_{HH}+S_{VV}{|}^2\\ T_{22}=\frac{1}{2}|S_{HH}-S_{VV}{|}^2\\ T_{33}=2|S_{HV}{|}^2\end{array}\right.---<3>$

Wherein, T₁₁、T₂₂、T₃₃For element on the diagonal of the coherence matrix T that polarizes.

3. the Classification of Polarimetric SAR Image method of non-down sampling contourlet convolutional neural networks according to claim 1, wherein carries out Pauli decomposition to filtered polarization scattering matrix S in step (2), and its step is as follows:

(2a) define basic collision matrix, be called Pauli base: { S_a,S_b,S_c,S_d, formula is as follows:

$\left\{\begin{array}{c}{S}_a=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right]\\ S_b=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& 0\\ 0& -1\end{array}\right]\\ S_c=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right]\\ S_d=\frac{1}{\sqrt{2}}\left[\begin{array}{cc}1& -i\\ i& 1\end{array}\right]\end{array}\right.---<4>$

Wherein $\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right]$ Represent odd scattering, $\left[\begin{array}{cc}1& 0\\ 0& -1\end{array}\right]$ Represent even scattering, $\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right]$ Represent volume scattering, $\left[\begin{array}{cc}0& -i\\ i& 0\end{array}\right]$ Representing non-existent type of ground objects, therefore d value is 0；

(2b) the Pauli base defined according to formula<4>, obtains the expression formula of polarization scattering matrix S:

$S=\left[\begin{array}{cc}{S}_{HH}& S_{HV}\\ S_{VH}& S_{VV}\end{array}\right]={\mathrm{aS}}_a+{\mathrm{bS}}_b+{\mathrm{cS}}_c+{\mathrm{dS}}_d---<5>$

The wherein value of a correspondence odd scattering, the value of b correspondence even scattering, c represents the value of volume scattering, and d represents the value of scattering composition corresponding to non-existent type of ground objects；

(2c) solve formula<5>, obtain scattering value a, b, c, d, be denoted as vector form as follows:

$K=\&lsqb;\begin{array}{cccc}a& b& c& d\end{array}\&rsqb;=\frac{1}{\sqrt{2}}{\left(\begin{array}{cccc}{S}_{HH}+S_{VV}& S_{HH}-S_{VV}& S_{HV}+S_{VH}& i\left(S_{VH}-S_{HV}\right)\end{array}\right)}^T---<6>$

When meeting reciprocity condition S_HV=S_VHTime, formula<6>is reduced to:

$K=\&lsqb;\begin{array}{ccc}a& b& c\end{array}\&rsqb;=\frac{1}{\sqrt{2}}\left(\begin{array}{ccc}{S}_{HH}+S_{VV}& S_{HH}-S_{VV}& 2S_{HV}\end{array}\right),---<7>$

The S that formula<3>is tried to achieve_HH、S_VV、S_HVSubstitution formula<7>, tries to achieve polarization characteristic K.

4. the Classification of Polarimetric SAR Image method of non-down sampling contourlet convolutional neural networks according to claim 1, wherein to the eigenmatrix F normalization based on pixel in step (3), adopt characteristic line pantography, namely first obtain the maximum max (F) of eigenmatrix F；Again by each element in eigenmatrix F all divided by maximum max (F), obtain normalized eigenmatrix F1.

5. the Classification of Polarimetric SAR Image method of non-down sampling contourlet convolutional neural networks according to claim 1, wherein the non-down sampling contourlet convolutional neural networks in step (6b), its structure is 8 layers, is expressed as:

Input layer → non-down sampling contourlet transform layer → pond layer → convolutional layer → pond layer → full articulamentum → full articulamentum → softmax grader

The parameter of each layer is:

1st layer of input layer: output characteristic mapping graph=3；

2nd layer of non-down sampling contourlet transform layer: output characteristic mapping graph=12；

3rd layer of pond layer: down-sampling yardstick=2.

4th layer of convolutional layer: output characteristic mapping graph=20, filter size=4；

5th layer of pond layer: down-sampling yardstick=2；

6th layer of full articulamentum: output characteristic mapping graph=100；

7th layer of full articulamentum: output characteristic mapping graph=64；

8th layer of softmax grader: output characteristic mapping graph=15.