!25 增加分水岭切图

Merge pull request !25 from 逐光/test

src/main/java/org/wlld/MatrixTools/Matrix.java

@@ -398,7 +398,6 @@ public class Matrix {
         if (this.x >= x && this.y >= y) {
             matrix[x][y] = number;
         } else {
-            System.out.println("myX==" + this.x + ",myY==" + this.y + ",x==" + x + ",y==" + y);
             throw new Exception("matrix length too little");
         }
     }

src/main/java/org/wlld/config/Kernel.java

@@ -8,6 +8,11 @@ public class Kernel {
     private static final String Horizontal_Number = "[-1,-2,-1]#[0,0,0]#[1,2,1]#";//横卷积核
     private static final String All_Number = "[1,-2,1]#[-2,4,-2]#[1,-2,1]#";//角卷积
     private static final String All_Number2 = "[-1,0,-1]#[0,4,0]#[-1,0,-1]#";
+    public static final int Region_Nub = 30;//一张图有多少份
+    public static final int Region_Dif = 3;//区域内的最大间隔
+    public static final double Region_Th = 0.2;//区域分割阈值
+    public static final double th = 0.5;//分水岭灰度阈值
+    public static final double rgbN = 441.6729559300637;//RGB范数归一化最大值
     public static Matrix Vertical;
     public static Matrix Horizontal;
     public static Matrix All;

src/main/java/org/wlld/imageRecognition/Convolution.java

@@ -106,8 +106,7 @@ public class Convolution extends Frequency {
         return features;
     }
 
-    public List<List<Double>> kc(ThreeChannelMatrix threeChannelMatrix, int poolSize, int sqNub
-            , int regionSize) throws Exception {
+    public List<Double> getCenterColor(ThreeChannelMatrix threeChannelMatrix, int poolSize, int sqNub) throws Exception {
         Matrix matrixR = threeChannelMatrix.getMatrixR();
         Matrix matrixG = threeChannelMatrix.getMatrixG();
         Matrix matrixB = threeChannelMatrix.getMatrixB();
@@ -117,7 +116,7 @@ public class Convolution extends Frequency {
         RGBSort rgbSort = new RGBSort();
         int x = matrixR.getX();
         int y = matrixR.getY();
-        meanClustering = new MeanClustering(sqNub);
+        MeanClustering meanClustering = new MeanClustering(sqNub);
         for (int i = 0; i < x; i++) {
             for (int j = 0; j < y; j++) {
                 double[] color = new double[]{matrixR.getNumber(i, j) / 255, matrixG.getNumber(i, j) / 255, matrixB.getNumber(i, j) / 255};
@@ -126,113 +125,52 @@ public class Convolution extends Frequency {
         }
         meanClustering.start();
         List<RGBNorm> rgbNorms = meanClustering.getMatrices();
-        double minNorm = 0;
-        int normSize = rgbNorms.size();
-        for (int i = 0; i < normSize; i++) {
-            RGBNorm rgbNorm = rgbNorms.get(i);
-            double[] rgb = rgbNorm.getRgb();
-            for (int j = 0; j < normSize; j++) {
-                if (j != i) {
-                    double normSub = getEDist(rgb, rgbNorms.get(j).getRgb());
-                    if (minNorm == 0 || normSub < minNorm) {
-                        minNorm = normSub;
-                    }
-                }
-            }
-        }
         Collections.sort(rgbNorms, rgbSort);
-        double[] features = new double[sqNub];
+        List<Double> feature = new ArrayList<>();
         for (int i = 0; i < sqNub; i++) {
-            features[i] = rgbNorms.get(i).getNorm();
+            feature.add(rgbNorms.get(i).getNorm());
         }
-       // System.out.println(Arrays.toString(features));
-        minNorm = ArithUtil.div(minNorm, 2);
-        return checkImage(matrixR, matrixG, matrixB, minNorm, regionSize, features);
+        return feature;
     }
 
-    private List<List<Double>> checkImage(Matrix matrixR, Matrix matrixG, Matrix matrixB, double minNorm, int size
-            , double[] features) throws Exception {
-        List<List<Double>> lists = new ArrayList<>();
-        int x = matrixR.getX() - size;//求导后矩阵的行数
-        int y = matrixR.getY() - size;//求导后矩阵的列数
-        for (int i = 0; i < x; i += size) {//遍历行
-            for (int j = 0; j < y; j += size) {//遍历每行的列
-                Matrix myMatrixR = matrixR.getSonOfMatrix(i, j, size, size);
-                Matrix myMatrixG = matrixG.getSonOfMatrix(i, j, size, size);
-                Matrix myMatrixB = matrixB.getSonOfMatrix(i, j, size, size);
-                List<Double> list = getListFeature(myMatrixR, myMatrixG, myMatrixB, minNorm, features);
-                // System.out.println("feature====" + list);
-                lists.add(list);
-            }
-        }
-        return lists;
-    }
-
-    private List<Double> getListFeature(Matrix matrixR, Matrix matrixG, Matrix matrixB, double minNorm
-            , double[] features) throws Exception {
-        int x = matrixR.getX();
-        int y = matrixR.getY();
-        List<Double> list = new ArrayList<>();
-        double[] feat = new double[features.length];
-        List<RGBNorm> rgbNormList = meanClustering.getMatrices();
-        for (int i = 0; i < x; i++) {
-            for (int j = 0; j < y; j++) {
-                double[] color = new double[]{matrixR.getNumber(i, j) / 255, matrixG.getNumber(i, j) / 255, matrixB.getNumber(i, j) / 255};
-                int id = -1;
-                double minDist = 0;
-                for (int t = 0; t < rgbNormList.size(); t++) {
-                    RGBNorm rgbNorm = rgbNormList.get(t);
-                    double dist = getEDist(color, rgbNorm.getRgb());
-                    if (minDist == 0 || dist < minDist) {
-                        minDist = dist;
-                        id = t;
-                    }
-                }
-                if (minDist >= minNorm) {
-                    id = -1;
-                }
-                if (id > -1) {
-                    feat[id] = rgbNormList.get(id).getNorm();
-                }
-            }
-        }
-        //
-        for (int i = 0; i < feat.length; i++) {
-            list.add(feat[i]);
-        }
-        return list;
+    private void regression(XYBody xyBody) {
+        //计算当前图形的线性回归
+        RegressionBody regressionBody = new RegressionBody();
+        regressionBody.lineRegression(xyBody.getY(), xyBody.getX(), this);
+        xyBody.setRegressionBody(regressionBody);
     }
 
-    public List<List<Double>> imageTrance(Matrix matrix, int size, int featureNub) throws Exception {//矩阵和卷积核大小
+    public XYBody imageTrance(Matrix matrix, int size) throws Exception {//矩阵和卷积核大小
         int xn = matrix.getX();
         int yn = matrix.getY();
         int xSize = xn / size;//求导后矩阵的行数
         int ySize = yn / size;//求导后矩阵的列数
         double[] Y = new double[xSize * ySize];
         double[] X = new double[xSize * ySize];
+        double rgbN = Kernel.rgbN;
         for (int i = 0; i < xn - size; i += size) {
             for (int j = 0; j < yn - size; j += size) {
                 Matrix matrix1 = matrix.getSonOfMatrix(i, j, size, size);
                 double[] nubs = new double[size * size];//平均值数组
                 for (int t = 0; t < size; t++) {
                     for (int k = 0; k < size; k++) {
-                        double nub = matrix1.getNumber(t, k) / 255;
+                        double nub = matrix1.getNumber(t, k) / rgbN;
                         nubs[t * size + k] = nub;
                     }
                 }
                 double avg = average(nubs);//平均值
-                double dc = dcByAvg(nubs, avg);//当前离散系数
-                //double va = varianceByAve(nubs, avg);//方差
+                //double dc = frequency.dcByAvg(nubs, avg);//当前离散系数
+                double va = varianceByAve(nubs, avg);//方差
                 //离散系数作为X，AVG作为Y
                 int t = i / size * ySize + j / size;
                 Y[t] = avg;
-                X[t] = dc;
+                X[t] = va;
             }
         }
-        //计算当前图形的线性回归
-        RegressionBody regressionBody = new RegressionBody();
-        regressionBody.lineRegression(Y, X, this);
-        return regressionBody.mappingMatrix(featureNub);
+        XYBody xyBody = new XYBody();
+        xyBody.setX(X);
+        xyBody.setY(Y);
+        return xyBody;
     }
 
 

src/main/java/org/wlld/imageRecognition/CoverBody.java

@@ -6,12 +6,21 @@ import java.util.Map;
 public class CoverBody {
     private List<List<Double>> feature;//特征
     private Map<Integer, Double> tag;//标注
+    private List<List<Double>> cFeature;//卷积特征
     private int type;
 
     public int getType() {
         return type;
     }
 
+    public List<List<Double>> getcFeature() {
+        return cFeature;
+    }
+
+    public void setcFeature(List<List<Double>> cFeature) {
+        this.cFeature = cFeature;
+    }
+
     public void setType(int type) {
         this.type = type;
     }

src/main/java/org/wlld/imageRecognition/MaxPoint.java

@@ -6,6 +6,20 @@ import org.wlld.i.OutBack;
 public class MaxPoint implements OutBack {
     private int id;
     private double point = -1;
+    private int pid = 0;//指定ID的概率
+    private double pPoint = 0;
+
+    public void setPid(int pid) {
+        this.pid = pid;
+    }
+
+    public double getpPoint() {
+        return pPoint;
+    }
+
+    public void setTh(double th) {
+        point = th;
+    }
 
     public int getId() {
         return id;
@@ -13,6 +27,9 @@ public class MaxPoint implements OutBack {
 
     @Override
     public void getBack(double out, int id, long eventId) {
+        if (pid > 0 && id == pid) {
+            pPoint = out;
+        }
         if (out > point) {
             point = out;
             this.id = id;

src/main/java/org/wlld/imageRecognition/Operation.java

@@ -10,6 +10,7 @@ import org.wlld.i.OutBack;
 import org.wlld.imageRecognition.border.*;
 import org.wlld.nerveCenter.NerveManager;
 import org.wlld.nerveCenter.Normalization;
+import org.wlld.nerveEntity.Nerve;
 import org.wlld.nerveEntity.SensoryNerve;
 import org.wlld.tools.ArithUtil;
 import org.wlld.tools.IdCreator;
@@ -285,6 +286,17 @@ public class Operation {//进行计算
         lvq.insertMatrixBody(matrixBody);
     }
 
+    private List<Double> getFeatures(Matrix matrix) throws Exception {
+        List<Double> list = new ArrayList<>();
+        for (int i = 0; i < matrix.getX(); i++) {
+            for (int j = 0; j < matrix.getY(); j++) {
+                double nub = matrix.getNumber(i, j);
+                list.add(nub);
+            }
+        }
+        return list;
+    }
+
     private List<Double> getFeature(Matrix matrix) throws Exception {//将特征矩阵转化为集合并除10
         List<Double> list = new ArrayList<>();
         Normalization normalization = templeConfig.getNormalization();
@@ -298,7 +310,6 @@ public class Operation {//进行计算
                 } else {
                     list.add(0.0);
                 }
-
             }
         }
         return list;
@@ -560,6 +571,17 @@ public class Operation {//进行计算
         return iou;
     }
 
+    public double toSeeById(Matrix matrix, int id) throws Exception {//返回某一个分类的概率
+        intoConvolutionNetwork(2, matrix, templeConfig.getConvolutionNerveManager().getSensoryNerves(),
+                false, 0, matrixBack);
+        List<Double> list = getFeature(matrixBack.getMatrix());
+        MaxPoint maxPoint = new MaxPoint();
+        maxPoint.setPid(id);
+        long t = IdCreator.get().nextId();
+        intoDnnNetwork(t, list, templeConfig.getSensoryNerves(), false, null, maxPoint);
+        return maxPoint.getpPoint();
+    }
+
     //图像视觉 Accuracy 模式
     public int toSee(Matrix matrix) throws Exception {
         if (templeConfig.getStudyPattern() == StudyPattern.Accuracy_Pattern) {
@@ -588,6 +610,7 @@ public class Operation {//进行计算
     private int getClassificationIdByDnn(Matrix myMatrix) throws Exception {
         List<Double> list = getFeature(myMatrix);
         MaxPoint maxPoint = new MaxPoint();
+        maxPoint.setTh(templeConfig.getTh());
         long id = IdCreator.get().nextId();
         intoDnnNetwork(id, list, templeConfig.getSensoryNerves(), false, null, maxPoint);
         return maxPoint.getId();
@@ -601,6 +624,7 @@ public class Operation {//进行计算
         for (Map.Entry<Integer, Matrix> entry : matrixK.entrySet()) {
             Matrix matrix = entry.getValue();
             double dist = MatrixOperation.getEDist(matrix, myVector);
+            System.out.println("距离===" + dist + ",类别==" + entry.getKey());
             if (minDist == 0 || dist < minDist) {
                 minDist = dist;
                 id = entry.getKey();

src/main/java/org/wlld/imageRecognition/Picture.java

@@ -77,16 +77,23 @@ public class Picture {
         ThreeChannelMatrix threeChannelMatrix = new ThreeChannelMatrix();
         Matrix matrixR = new Matrix(height, width);//行，列
         Matrix matrixG = new Matrix(height, width);//行，列
-        Matrix matrixB = new Matrix(height, width);//行，列
+        Matrix matrixB = new Matrix(height, width);//行，
+        Matrix matrixRGB = new Matrix(height, width);
         threeChannelMatrix.setMatrixR(matrixR);
         threeChannelMatrix.setMatrixG(matrixG);
         threeChannelMatrix.setMatrixB(matrixB);
+        threeChannelMatrix.setMatrixRGB(matrixRGB);
         for (int i = 0; i < height; i++) {
             for (int j = 0; j < width; j++) {
                 int pixel = bi.getRGB(j, i);// 下面三行代码将一个数字转换为RGB数字
-                matrixR.setNub(i, j, (pixel & 0xff0000) >> 16);
-                matrixG.setNub(i, j, (pixel & 0xff00) >> 8);
-                matrixB.setNub(i, j, (pixel & 0xff));
+                int r = (pixel & 0xff0000) >> 16;//R
+                int g = (pixel & 0xff00) >> 8;//G
+                int b = (pixel & 0xff);//B
+                double rgb = Math.sqrt(Math.pow(r, 2) + Math.pow(g, 2) + Math.pow(b, 2));
+                matrixRGB.setNub(i, j, rgb);
+                matrixR.setNub(i, j, r);
+                matrixG.setNub(i, j, g);
+                matrixB.setNub(i, j, b);
             }
         }
         return threeChannelMatrix;

src/main/java/org/wlld/imageRecognition/TempleConfig.java

@@ -41,7 +41,7 @@ public class TempleConfig {
     private boolean isHavePosition = false;//是否需要锁定物体位置
     private LVQ lvq;//模型需要返回,精准模式下的原型聚类
     private Frame frame;//先验边框
-    private double th = 0.6;//标准阈值
+    private double th = -1;//标准阈值
     private boolean boxReady = false;//边框已经学习完毕
     private double iouTh = 0.5;//IOU阈值
     private int lvqNub = 10;//lvq循环次数，默认30
@@ -255,6 +255,31 @@ public class TempleConfig {
         }
     }
 
+    public void initCover(boolean initPower, int width, int height, int kNub) throws Exception {
+        int deep = 0;
+        Map<Integer, Matrix> matrixMap = new HashMap<>();//主键与期望矩阵的映射
+        while (width > 5 && height > 5) {
+            width = width / 3;
+            height = height / 3;
+            deep++;
+        }
+        int nub = height * width * 3 + kNub;
+        initNerveManager(true, nub, this.deep, studyPoint);
+        //加载各识别分类的期望矩阵
+        matrixMap.put(0, new Matrix(height, width));
+        for (int k = 1; k <= classificationNub; k++) {
+            Matrix matrix = new Matrix(height, width);//初始化期望矩阵
+            double t = k * matrixWidth;//期望矩阵的分类参数数值
+            for (int i = 0; i < height; i++) {//给期望矩阵注入期望参数
+                for (int j = 0; j < width; j++) {
+                    matrix.setNub(i, j, t);
+                }
+            }
+            matrixMap.put(k, matrix);
+        }
+        convolutionNerveManager = initNerveManager(matrixMap, initPower, deep);
+    }
+
     private void initModelVision(boolean initPower, int width, int height) throws Exception {//初始标准模板视觉
         double d;
         if (width > height) {

src/main/java/org/wlld/imageRecognition/ThreeChannelMatrix.java

@@ -6,6 +6,24 @@ public class ThreeChannelMatrix {
     Matrix matrixR;
     Matrix matrixG;
     Matrix matrixB;
+    Matrix H;
+    Matrix matrixRGB;
+
+    public Matrix getMatrixRGB() {
+        return matrixRGB;
+    }
+
+    public void setMatrixRGB(Matrix matrixRGB) {
+        this.matrixRGB = matrixRGB;
+    }
+
+    public Matrix getH() {
+        return H;
+    }
+
+    public void setH(Matrix h) {
+        H = h;
+    }
 
     public Matrix getMatrixR() {
         return matrixR;

src/main/java/org/wlld/imageRecognition/XYBody.java

@@ -0,0 +1,33 @@
+package org.wlld.imageRecognition;
+
+import org.wlld.imageRecognition.modelEntity.RegressionBody;
+
+public class XYBody {
+    private double[] X;
+    private double[] Y;
+    private RegressionBody regressionBody;
+
+    public RegressionBody getRegressionBody() {
+        return regressionBody;
+    }
+
+    public void setRegressionBody(RegressionBody regressionBody) {
+        this.regressionBody = regressionBody;
+    }
+
+    public double[] getX() {
+        return X;
+    }
+
+    public void setX(double[] x) {
+        X = x;
+    }
+
+    public double[] getY() {
+        return Y;
+    }
+
+    public void setY(double[] y) {
+        Y = y;
+    }
+}

src/main/java/org/wlld/imageRecognition/border/LVQ.java

@@ -16,7 +16,7 @@ public class LVQ {
     private int typeNub;//原型聚类个数,即分类个数(需要模型返回)
     private MatrixBody[] model;//原型向量(需要模型返回)
     private List<MatrixBody> matrixList = new ArrayList<>();
-    private double studyPoint = 0.1;//量化学习率
+    private double studyPoint = 0.01;//量化学习率
     private int length;//向量长度(需要返回)
     private boolean isReady = false;
     private int lvqNub;

src/main/java/org/wlld/imageRecognition/modelEntity/RegressionBody.java

@@ -9,26 +9,26 @@ import java.util.List;
 public class RegressionBody {
     private double w;
     private double b;
-    private double[] X;
+    private double maxDis;//最大距离
 
-    public List<List<Double>> mappingMatrix(int size) {
-        int len = X.length - size;
-        List<List<Double>> lists = new ArrayList<>();
-        for (int i = 0; i < len; i += size) {
-            List<Double> list = new ArrayList<>();
-            for (int t = i; t < i + size; t++) {
-                double nub = ArithUtil.add(ArithUtil.mul(w, X[t]), b);
-                list.add(nub);
+    public void getMaxDis(double[] Y, double[] X) {//获取当前的最大距离
+        double allNub = 0;
+        for (int i = 0; i < X.length; i++) {
+            double y = ArithUtil.add(ArithUtil.mul(X[i], w), b);
+            double dis = Math.abs(ArithUtil.sub(Y[i], y));
+            allNub = ArithUtil.add(allNub, dis);
+            System.out.println("dis======" + dis);
+            if (dis > maxDis) {
+                maxDis = dis;
             }
-            lists.add(list);
         }
-        return lists;
+        allNub = ArithUtil.div(allNub, X.length);//当前最大值：0.1955576793420405,allNub==0.035958733
+        System.out.println("当前最大值：" + maxDis + ",allNub==" + allNub);
     }
 
     public void lineRegression(double[] Y, double[] X, Frequency frequency) {//进行二元线性回归
         double avX = frequency.average(X);//平均值
         int len = Y.length;
-        this.X = X;
         double wFenZi = 0;
         double wFenMu;
         double xSigma = 0;//X求和
@@ -49,6 +49,5 @@ public class RegressionBody {
             bSigma = ArithUtil.add(ArithUtil.sub(y, ArithUtil.mul(w, x)), bSigma);
         }
         b = ArithUtil.div(bSigma, len);
-
     }
 }

src/main/java/org/wlld/imageRecognition/segmentation/ImageSegmentation.java

@@ -1,49 +0,0 @@
-package org.wlld.imageRecognition.segmentation;
-
-import org.wlld.MatrixTools.Matrix;
-import org.wlld.imageRecognition.ThreeChannelMatrix;
-
-import java.util.HashMap;
-import java.util.Map;
-
-/**
- * @author lidapeng
- * @description 阈值分区
- * @date 10:25 上午 2020/1/13
- */
-public class ImageSegmentation {
-    private Matrix matrixR;
-    private Matrix matrixG;
-    private Matrix matrixB;
-    private Map<Integer, RegionBody> regionBodyList = new HashMap<>();
-    private int id = 1;
-
-    public ImageSegmentation(ThreeChannelMatrix threeChannelMatrix) throws Exception {
-        matrixR = threeChannelMatrix.getMatrixR();
-        matrixG = threeChannelMatrix.getMatrixG();
-        matrixB = threeChannelMatrix.getMatrixB();
-        int x = matrixR.getX();
-        int y = matrixR.getY();
-        for (int i = 0; i < x; i++) {
-            for (int j = 0; j < y; j++) {
-                double r = matrixR.getNumber(i, j);
-                double g = matrixG.getNumber(i, j);
-                double b = matrixB.getNumber(i, j);
-
-            }
-        }
-    }
-
-    public int getMin(double[] array) {
-        double min = array[0];
-        int minIdx = 0;
-        for (int i = 1; i < array.length; i++) {
-            if (array[i] < min) {
-                minIdx = i;
-                min = array[i];
-            }
-        }
-        return minIdx;
-    }
-
-}

src/main/java/org/wlld/imageRecognition/segmentation/RegionBody.java

@@ -1,14 +1,46 @@
 package org.wlld.imageRecognition.segmentation;
 
-import org.wlld.MatrixTools.Matrix;
-
-import java.util.ArrayList;
-import java.util.List;
-
 /**
  * @author lidapeng
  * @description 分区实体
  */
 public class RegionBody {
+    private int minX = -1;
+    private int minY = -1;
+    private int maxX;
+    private int maxY;
+
+    public void setX(int x) {
+        if (x < minX || minX == -1) {
+            minX = x;
+        }
+        if (x > maxX) {
+            maxX = x;
+        }
+    }
+
+    public void setY(int y) {
+        if (y < minY || minY == -1) {
+            minY = y;
+        }
+        if (y > maxY) {
+            maxY = y;
+        }
+    }
+
+    public int getMinX() {
+        return minX;
+    }
+
+    public int getMinY() {
+        return minY;
+    }
+
+    public int getMaxX() {
+        return maxX;
+    }
 
+    public int getMaxY() {
+        return maxY;
+    }
 }

src/main/java/org/wlld/imageRecognition/segmentation/Specifications.java

@@ -0,0 +1,22 @@
+package org.wlld.imageRecognition.segmentation;
+
+public class Specifications {
+    private double width;
+    private double height;
+
+    public double getWidth() {
+        return width;
+    }
+
+    public void setWidth(double width) {
+        this.width = width;
+    }
+
+    public double getHeight() {
+        return height;
+    }
+
+    public void setHeight(double height) {
+        this.height = height;
+    }
+}

src/main/java/org/wlld/tools/Frequency.java

@@ -14,7 +14,7 @@ public abstract class Frequency {//统计频数
         return allNub;
     }
 
-    public double getEDist(double[] x1, double[] x2) {//返回两个等长数组之间的欧式距离
+    public static double getEDist(double[] x1, double[] x2) {//返回两个等长数组之间的欧式距离
         double[] y = new double[x1.length];
         for (int i = 0; i < y.length; i++) {
             y[i] = x1[i] - x2[i];

src/test/java/coverTest/CoverTest.java

@@ -4,6 +4,7 @@ import com.alibaba.fastjson.JSON;
 import com.alibaba.fastjson.JSONObject;
 import org.wlld.MatrixTools.Matrix;
 import org.wlld.ModelData;
+import org.wlld.config.Classifier;
 import org.wlld.config.RZ;
 import org.wlld.config.StudyPattern;
 import org.wlld.function.Sigmod;
@@ -22,6 +23,18 @@ import java.util.Map;
  */
 public class CoverTest {
     public static void main(String[] args) throws Exception {
+//        TempleConfig templeConfig = new TempleConfig();
+//        templeConfig.setClassifier(Classifier.DNN);
+//        templeConfig.setTh(0);
+//        templeConfig.setSoftMax(true);
+//        templeConfig.init(StudyPattern.Accuracy_Pattern, false, 640, 480, 2);
+//        ModelParameter modelParameter = JSON.parseObject(ModelData.DATA3, ModelParameter.class);
+//        templeConfig.insertModel(modelParameter);
+//        Operation operation = new Operation(templeConfig);
+//        for (int i = 1; i < 100; i++) {
+//            cover2(operation, "D:\\pic\\6\\a/a" + i + ".jpg");
+//        }
+        //test(null, 2, 3, 40, "c", 3);
         cover();
     }
 
@@ -40,58 +53,51 @@ public class CoverTest {
     public static Operation getModel() throws Exception {
         //覆盖率计算,计算以前，内存中已经注入过模型了
         TempleConfig templeConfig = new TempleConfig();
-        templeConfig.init(StudyPattern.Cover_Pattern, true, 400, 400, 3);
+        templeConfig.setSensoryNerveNub(3);
+        templeConfig.setSoftMax(true);
+        templeConfig.setDeep(2);
+        templeConfig.setHiddenNerveNub(9);
+        templeConfig.init(StudyPattern.Cover_Pattern, true, 400, 400, 4);
         ModelParameter modelParameter = JSONObject.parseObject(ModelData.DATA, ModelParameter.class);
         templeConfig.insertModel(modelParameter);
         return new Operation(templeConfig);//初始化运算类
     }
 
-    public static void test(Operation operation, int poolSize, int sqlNub, int regionSize) throws Exception {
+    public static void test(Operation operation, int poolSize, int sqlNub, int regionSize,
+                            String name, int t) throws Exception {
         Picture picture = new Picture();
-        int allNub = 0;
         int wrong = 0;
-        for (int i = 1; i < 2; i++) {
-            allNub += 3;
-            ThreeChannelMatrix threeChannelMatrix = picture.getThreeMatrix("D:\\share\\jie/1.jpg");
-            ThreeChannelMatrix threeChannelMatrix2 = picture.getThreeMatrix("D:\\share\\jie/2.jpg");
-         //   ThreeChannelMatrix threeChannelMatrix3 = picture.getThreeMatrix("D:\\share\\jie/3.jpg");
-            //ThreeChannelMatrix threeChannelMatrix4 = picture.getThreeMatrix("D:\\share\\cai/d" + i + ".jpg");
-            //ThreeChannelMatrix threeChannelMatrix5 = picture.getThreeMatrix("D:\\share\\cai/e" + i + ".jpg");
-            //ThreeChannelMatrix threeChannelMatrix6 = picture.getThreeMatrix("D:\\share\\cai/f" + i + ".jpg");
-            //ThreeChannelMatrix threeChannelMatrix7 = picture.getThreeMatrix("D:\\share\\cai/g" + i + ".jpg");
-            //ThreeChannelMatrix threeChannelMatrix8 = picture.getThreeMatrix("D:\\share\\cai/h" + i + ".jpg");
-
+        if (operation == null) {
+            operation = getModel();
+        }
+        for (int i = 1; i < 100; i++) {
+            String na = "D:\\pic\\test/" + name + i + ".jpg";
+            //System.out.println("name======================" + na);
+            ThreeChannelMatrix threeChannelMatrix = picture.getThreeMatrix(na);
             Map<Integer, Double> map1 = operation.coverPoint(threeChannelMatrix, poolSize, sqlNub, regionSize);
-            Map<Integer, Double> map2 = operation.coverPoint(threeChannelMatrix2, poolSize, sqlNub, regionSize);
-           // Map<Integer, Double> map3 = operation.coverPoint  (threeChannelMatrix3, poolSize, sqlNub, regionSize);
-            //Map<Integer, Double> map4 = operation.coverPoint(threeChannelMatrix4, poolSize, sqlNub, regionSize);
-//            Map<Integer, Double> map5 = operation.coverPoint(threeChannelMatrix5, poolSize, sqlNub, regionSize);
-//            Map<Integer, Double> map6 = operation.coverPoint(threeChannelMatrix6, poolSize, sqlNub, regionSize);
-//            Map<Integer, Double> map7 = operation.coverPoint(threeChannelMatrix7, poolSize, sqlNub, regionSize);
-//            Map<Integer, Double> map8 = operation.coverPoint(threeChannelMatrix8, poolSize, sqlNub, regionSize);
-
+            int id = 0;
+            double point = 0;
             for (Map.Entry<Integer, Double> entry : map1.entrySet()) {
                 int key = entry.getKey();
                 double value = entry.getValue();
-                System.out.println("1key===" + key + ",value==" + value);
+                // System.out.println("key==" + key + ",value==" + value);
+                if (value > point) {
+                    point = value;
+                    id = key;
+                }
             }
-            System.out.println("=============================");
-            for (Map.Entry<Integer, Double> entry : map2.entrySet()) {
-                int key = entry.getKey();
-                double value = entry.getValue();
-                System.out.println("2key===" + key + ",value==" + value);
+            if (id != t) {
+                wrong++;
             }
-            System.out.println("=============================");
-
-//            for (Map.Entry<Integer, Double> entry : map3.entrySet()) {
-//                int key = entry.getKey();
-//                double value = entry.getValue();
-//                System.out.println("3key===" + key + ",value==" + value);
-//            }
-//            System.out.println("=============================");
-
         }
+        System.out.println("错误率：" + wrong + "%");
+    }
 
+    public static void cover2(Operation operation, String url) throws Exception {
+        Picture picture = new Picture();
+        Matrix matrix = picture.getImageMatrixByLocal(url);
+        double type = operation.toSeeById(matrix, 1);
+        System.out.println("type==" + type);
     }
 
     public static void cover() throws Exception {
@@ -101,39 +107,27 @@ public class CoverTest {
         TempleConfig templeConfig = new TempleConfig();
         //初始化模板 注意 width height参数是你训练图片的实际尺寸需要改，其他不用动
         //创建运算类进行标注
-        //templeConfig.setActiveFunction(new Sigmod());
         templeConfig.isShowLog(true);
         templeConfig.setStudyPoint(0.01);//不动
         templeConfig.setSoftMax(true);
         //templeConfig.setDeep(2);
-        templeConfig.setSensoryNerveNub(2);
-        templeConfig.setRzType(RZ.L1);//不动
+        //templeConfig.setHiddenNerveNub(9);
+        templeConfig.setSensoryNerveNub(3);//多出来的
+        templeConfig.setRzType(RZ.L1);//不动//3 18
         templeConfig.setlParam(0.015);//不动
         templeConfig.init(StudyPattern.Cover_Pattern, true, 400, 400, 2);
         Operation operation = new Operation(templeConfig);
-        for (int i = 1; i < 2; i++) {
+        for (int i = 1; i < 100; i++) {
             Map<Integer, ThreeChannelMatrix> matrixMap = new HashMap<>();
-            ThreeChannelMatrix threeChannelMatrix1 = picture.getThreeMatrix("D:\\share\\jie/1.jpg");
-            ThreeChannelMatrix threeChannelMatrix2 = picture.getThreeMatrix("D:\\share\\jie/2.jpg");
-            // ThreeChannelMatrix threeChannelMatrix3 = picture.getThreeMatrix("D:\\share\\jie/3.jpg");
-//            ThreeChannelMatrix threeChannelMatrix4 = picture.getThreeMatrix("D:\\share\\cai/d" + i + ".jpg");
-//            ThreeChannelMatrix threeChannelMatrix5 = picture.getThreeMatrix("D:\\share\\cai/e" + i + ".jpg");
-//            ThreeChannelMatrix threeChannelMatrix6 = picture.getThreeMatrix("D:\\share\\cai/f" + i + ".jpg");
-//            ThreeChannelMatrix threeChannelMatrix7 = picture.getThreeMatrix("D:\\share\\cai/g" + i + ".jpg");
-//            ThreeChannelMatrix threeChannelMatrix8 = picture.getThreeMatrix("D:\\share\\cai/h" + i + ".jpg");
-            matrixMap.put(1, threeChannelMatrix1);
-            matrixMap.put(2, threeChannelMatrix2);
-            // matrixMap.put(3, threeChannelMatrix3);
-//            matrixMap.put(4, threeChannelMatrix4);
-//            matrixMap.put(5, threeChannelMatrix5);
-//            matrixMap.put(6, threeChannelMatrix6);
-//            matrixMap.put(7, threeChannelMatrix7);
-//            matrixMap.put(8, threeChannelMatrix8);
-            operation.coverStudy(matrixMap, 2, 2, 40, 20);
+            ThreeChannelMatrix threeChannelMatrix1 = picture.getThreeMatrix("D:\\pic\\test/b" + i + ".jpg");
+            ThreeChannelMatrix threeChannelMatrix2 = picture.getThreeMatrix("D:\\pic\\test/d" + i + ".jpg");
+            matrixMap.put(1, threeChannelMatrix1);//桔梗覆盖
+            matrixMap.put(2, threeChannelMatrix2);//土地
+            operation.coverStudy(matrixMap, 2, 3, 18, 2);
         }
         ModelParameter modelParameter = templeConfig.getModel();
         String model = JSON.toJSONString(modelParameter);
         System.out.println(model);
-        test(operation, 2, 2, 40);
+        test(operation, 2, 3, 18, "d", 2);
     }
 }

src/test/java/coverTest/FoodTest.java

@@ -2,119 +2,116 @@ package coverTest;
 
 import com.alibaba.fastjson.JSON;
 import org.wlld.MatrixTools.Matrix;
-import org.wlld.ModelData;
 import org.wlld.config.Classifier;
 import org.wlld.config.RZ;
 import org.wlld.config.StudyPattern;
 import org.wlld.imageRecognition.*;
-import org.wlld.imageRecognition.segmentation.ImageSegmentation;
+import org.wlld.imageRecognition.segmentation.RegionBody;
+import org.wlld.imageRecognition.segmentation.Specifications;
+import org.wlld.imageRecognition.segmentation.Watershed;
+import org.wlld.nerveCenter.NerveManager;
 import org.wlld.nerveEntity.ModelParameter;
 import org.wlld.tools.ArithUtil;
 
-import java.util.*;
+import java.util.ArrayList;
+import java.util.List;
 
 public class FoodTest {
 
     public static void main(String[] args) throws Exception {
+        //food();
+        rain();
+    }
+
+    public static void rain() throws Exception {//降雨
+        Convolution convolution = new Convolution();
+        Picture picture = new Picture();
+        Matrix matrix = picture.getImageMatrixByLocal("D:\\share/c.png");
+        List<Specifications> specificationsList = new ArrayList<>();
+        Specifications specifications = new Specifications();
+        specifications.setWidth(400);
+        specifications.setHeight(400);
+        specificationsList.add(specifications);
+        Watershed watershed = new Watershed(matrix, specificationsList);
+        List<RegionBody> regionList = watershed.rainfall();
 
-        food();
     }
 
-    public static void food() throws Exception {
+    public static void study() throws Exception {
         Picture picture = new Picture();
-        TempleConfig templeConfig = new TempleConfig();
+        Convolution convolution = new Convolution();
+        ThreeChannelMatrix threeChannelMatrix = picture.getThreeMatrix("");
+        List<Double> feature = convolution.getCenterColor(threeChannelMatrix, 2, 4);
+
+    }
 
-        templeConfig.setClassifier(Classifier.DNN);
-        templeConfig.isShowLog(true);
-        templeConfig.setRzType(RZ.L2);
-        templeConfig.setlParam(0.01);//0.015
-        templeConfig.init(StudyPattern.Accuracy_Pattern, true, 400, 400, 3);
-//      ModelParameter modelParameter1 = JSON.parseObject(ModelData.DATA, ModelParameter.class);
-//      templeConfig.insertModel(modelParameter1);
-        Operation operation = new Operation(templeConfig);
+    public static void food() throws Exception {
+        Picture picture = new Picture();//创建图片解析类
+        TempleConfig templeConfig = new TempleConfig();//创建配置模板类
+        templeConfig.setClassifier(Classifier.DNN);//使用DNN 分类器
+        //templeConfig.setActiveFunction(new Sigmod());//设置激活函数
+        templeConfig.setDeep(2);//设置深度 深度神经网络 深度越深速度越慢
+        //数量越大越准 但是影响量比较小 不绝对 盲试
+        templeConfig.setHiddenNerveNub(9);//设置隐层神经元数量
+        templeConfig.isShowLog(true);//输出打印数据
+        //
+        //templeConfig.setSoftMax(true);//启用最后一层的SOFTMAX
+        //templeConfig.setTh(-1);//设置阈值
+        templeConfig.setStudyPoint(0.012);//设置学习率 0-1
+        templeConfig.setRzType(RZ.L1);//设置正则函数
+        templeConfig.setlParam(0.015);//设置正则参数
+
+        templeConfig.init(StudyPattern.Accuracy_Pattern, true, 640, 480, 2);
+        Operation operation = new Operation(templeConfig);//计算类
         // 一阶段
         for (int j = 0; j < 2; j++) {
             for (int i = 1; i < 101; i++) {//一阶段
                 System.out.println("study1===================" + i);
                 //读取本地URL地址图片,并转化成矩阵
-                Matrix a = picture.getImageMatrixByLocal("D:\\share\\cai/a" + i + ".jpg");
-                Matrix b = picture.getImageMatrixByLocal("D:\\share\\cai/b" + i + ".jpg");
-                Matrix c = picture.getImageMatrixByLocal("D:\\share\\cai/c" + i + ".jpg");
-                //Matrix d = picture.getImageMatrixByLocal("D:\\share\\cai/d" + i + ".jpg");
-                //将图像矩阵和标注加入进行学习，Accuracy_Pattern 模式 进行第二次学习
-                //第二次学习的时候，第三个参数必须是 true
+                Matrix a = picture.getImageMatrixByLocal("D:\\pic\\1/a" + i + ".jpg");
+                Matrix b = picture.getImageMatrixByLocal("D:\\pic\\2/b" + i + ".jpg");
                 operation.learning(a, 1, false);
                 operation.learning(b, 2, false);
-                operation.learning(c, 3, false);
-                //operation.learning(d, 4, false);
             }
         }
-        //二阶段
+        // 二阶段 归一化
         for (int i = 1; i < 101; i++) {
             System.out.println("avg==" + i);
-            Matrix a = picture.getImageMatrixByLocal("D:\\share\\cai/a" + i + ".jpg");
-            Matrix b = picture.getImageMatrixByLocal("D:\\share\\cai/b" + i + ".jpg");
-            Matrix c = picture.getImageMatrixByLocal("D:\\share\\cai/c" + i + ".jpg");
-            //Matrix d = picture.getImageMatrixByLocal("D:\\share\\picture/d" + i + ".jpg");
+            Matrix a = picture.getImageMatrixByLocal("D:\\pic\\1/a" + i + ".jpg");
+            Matrix b = picture.getImageMatrixByLocal("D:\\pic\\2/b" + i + ".jpg");
             operation.normalization(a, templeConfig.getConvolutionNerveManager());
             operation.normalization(b, templeConfig.getConvolutionNerveManager());
-            operation.normalization(c, templeConfig.getConvolutionNerveManager());
-            //operation.normalization(d, templeConfig.getConvolutionNerveManager());
         }
         templeConfig.getNormalization().avg();
-//        ModelParameter modelParameter2 = templeConfig.getModel();
-//        String model2 = JSON.toJSONString(modelParameter2);
-//        System.out.println(model2);
-        for (int j = 0; j < 10; j++) {
+
+        for (int j = 0; j < 3; j++) {
             for (int i = 1; i < 101; i++) {
                 System.out.println("j==" + j + ",study2==================" + i);
-                //读取本地URL地址图片,并转化成矩阵
-                Matrix a = picture.getImageMatrixByLocal("D:\\share\\cai/a" + i + ".jpg");
-                Matrix b = picture.getImageMatrixByLocal("D:\\share\\cai/b" + i + ".jpg");
-                Matrix c = picture.getImageMatrixByLocal("D:\\share\\cai/c" + i + ".jpg");
-                // Matrix d = picture.getImageMatrixByLocal("D:\\share\\picture/d" + i + ".jpg");
-                //将图像矩阵和标注加入进行学习，Accuracy_Pattern 模式 进行第二次学习
-                //第二次学习的时候，第三个参数必须是 true
+                Matrix a = picture.getImageMatrixByLocal("D:\\pic\\1/a" + i + ".jpg");
+                Matrix b = picture.getImageMatrixByLocal("D:\\pic\\2/b" + i + ".jpg");
                 operation.learning(a, 1, true);
                 operation.learning(b, 2, true);
-                operation.learning(c, 3, true);
-                //operation.learning(d, 4, true);
             }
         }
         templeConfig.finishStudy();//结束学习
+        ModelParameter modelParameter = templeConfig.getModel();
+        String model = JSON.toJSONString(modelParameter);
+        System.out.println(model);
 
         int wrong = 0;
         int allNub = 0;
-        for (int i = 10; i <= 20; i++) {
+        for (int i = 1; i <= 100; i++) {
             //读取本地URL地址图片,并转化成矩阵
-            Matrix a = picture.getImageMatrixByLocal("D:\\share\\cai/a" + i + ".jpg");
-            Matrix b = picture.getImageMatrixByLocal("D:\\share\\cai/b" + i + ".jpg");
-            Matrix c = picture.getImageMatrixByLocal("D:\\share\\cai/c" + i + ".jpg");
-            //Matrix d = picture.getImageMatrixByLocal("D:\\share\\picture/d" + i + ".jpg");
-            //将图像矩阵和标注加入进行学习，Accuracy_Pattern 模式 进行第二次学习
-            //第二次学习的时候，第三个参数必须是 true
-            allNub += 3;
+            Matrix a = picture.getImageMatrixByLocal("D:\\pic\\1/a" + i + ".jpg");
+            allNub++;
             int an = operation.toSee(a);
             if (an != 1) {
                 wrong++;
             }
-            int bn = operation.toSee(b);
-            if (bn != 2) {
-                wrong++;
-            }
-            int cn = operation.toSee(c);
-            if (cn != 3) {
-                wrong++;
-            }
-            /*int dn = operation.toSee(d);
-            if (dn != 4) {
-                wrong++;
-            }*/
         }
         double wrongPoint = ArithUtil.div(wrong, allNub);
         System.out.println("错误率：" + (wrongPoint * 100) + "%");
-        ModelParameter modelParameter = templeConfig.getModel();
-        String model = JSON.toJSONString(modelParameter);
-        System.out.println(model);
+
     }
+
 }

src/test/java/org/wlld/HelloWorld.java

@@ -41,7 +41,6 @@ public class HelloWorld {
 
     public static void pictureDemo1() throws Exception {//图像学习DEMO
         //easyAI 包持续更新，现阶段一直在优化
-        //
         Picture picture = new Picture();
         //使用精度计算
         TempleConfig templeConfig = new TempleConfig(false, true);