1、OCR：新增表格识别模型

2、OCR：新增9个通用模型
3、OCR：支持批量检测识别
4、OCR：新增更多参数，使用更加灵活
5、人脸识别：支持ID查询及分页获取人脸信息
6、活体检测：视频检测支持设置最大帧数

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/utils/ConvertHtml2Excel.java

@@ -0,0 +1,233 @@
+package cn.smartjavaai.ocr.utils;
+
+import org.apache.commons.lang3.StringUtils;
+import org.apache.commons.lang3.math.NumberUtils;
+import org.apache.poi.hssf.usermodel.*;
+import org.apache.poi.ss.usermodel.*;
+import org.apache.poi.ss.util.CellRangeAddress;
+import org.dom4j.Document;
+import org.dom4j.DocumentException;
+import org.dom4j.DocumentHelper;
+import org.dom4j.Element;
+
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * @Auther: xiaoqiang
+ * @Date: 2020/12/9 9:16
+ * @Description:
+ */
+public class ConvertHtml2Excel {
+
+    /**
+     * html表格转excel
+     *
+     * @param tableHtml 如
+     *            <table>
+     *            ..
+     *            </table>
+     * @return
+     */
+    public static HSSFWorkbook table2Excel(String tableHtml) {
+        HSSFWorkbook wb = new HSSFWorkbook();
+        HSSFSheet sheet = wb.createSheet();
+        List<CrossRangeCellMeta> crossRowEleMetaLs = new ArrayList<>();
+        int rowIndex = 0;
+        try {
+            Document data = DocumentHelper.parseText(tableHtml);
+            // 生成表头
+            Element thead = data.getRootElement().element("thead");
+            HSSFCellStyle titleStyle = getTitleStyle(wb);
+            int ls=0;//列数
+            if (thead != null) {
+                List<Element> trLs = thead.elements("tr");
+                for (Element trEle : trLs) {
+                    HSSFRow row = sheet.createRow(rowIndex);
+                    List<Element> thLs = trEle.elements("td");
+                    ls=thLs.size();
+                    makeRowCell(thLs, rowIndex, row, 0, titleStyle, crossRowEleMetaLs);
+                    rowIndex++;
+                }
+            }
+            // 生成表体
+            Element tbody = data.getRootElement().element("tbody");
+            HSSFCellStyle contentStyle = getContentStyle(wb);
+            if (tbody != null) {
+                List<Element> trLs = tbody.elements("tr");
+                for (Element trEle : trLs) {
+                    HSSFRow row = sheet.createRow(rowIndex);
+                    List<Element> thLs = trEle.elements("th");
+                    int cellIndex = makeRowCell(thLs, rowIndex, row, 0, titleStyle, crossRowEleMetaLs);
+                    List<Element> tdLs = trEle.elements("td");
+                    makeRowCell(tdLs, rowIndex, row, cellIndex, contentStyle, crossRowEleMetaLs);
+                    rowIndex++;
+                }
+            }
+            // 合并表头
+            for (CrossRangeCellMeta crcm : crossRowEleMetaLs) {
+                sheet.addMergedRegion(new CellRangeAddress(crcm.getFirstRow(), crcm.getLastRow(), crcm.getFirstCol(), crcm.getLastCol()));
+                setRegionStyle(sheet, new CellRangeAddress(crcm.getFirstRow(), crcm.getLastRow(), crcm.getFirstCol(), crcm.getLastCol()),titleStyle);
+            }
+            for(int i=0;i<sheet.getRow(0).getPhysicalNumberOfCells();i++){
+                sheet.autoSizeColumn(i, true);//设置列宽
+                if(sheet.getColumnWidth(i)<255*256){
+                    sheet.setColumnWidth(i, sheet.getColumnWidth(i) < 9000 ? 9000 : sheet.getColumnWidth(i));
+                }else{
+                    sheet.setColumnWidth(i, 15000);
+                }
+            }
+        } catch (DocumentException e) {
+            e.printStackTrace();
+        }
+
+        return wb;
+    }
+
+    /**
+     * 生产行内容
+     *
+     * @return 最后一列的cell index
+     */
+    /**
+     * @param tdLs th或者td集合
+     * @param rowIndex 行号
+     * @param row POI行对象
+     * @param startCellIndex
+     * @param cellStyle 样式
+     * @param crossRowEleMetaLs 跨行元数据集合
+     * @return
+     */
+    private static int makeRowCell(List<Element> tdLs, int rowIndex, HSSFRow row, int startCellIndex, HSSFCellStyle cellStyle,
+                                   List<CrossRangeCellMeta> crossRowEleMetaLs) {
+        int i = startCellIndex;
+        for (int eleIndex = 0; eleIndex < tdLs.size(); i++, eleIndex++) {
+            int captureCellSize = getCaptureCellSize(rowIndex, i, crossRowEleMetaLs);
+            while (captureCellSize > 0) {
+                for (int j = 0; j < captureCellSize; j++) {// 当前行跨列处理（补单元格）
+                    row.createCell(i);
+                    i++;
+                }
+                captureCellSize = getCaptureCellSize(rowIndex, i, crossRowEleMetaLs);
+            }
+            Element thEle = tdLs.get(eleIndex);
+            String val = thEle.getTextTrim();
+            if (StringUtils.isBlank(val)) {
+                Element e = thEle.element("a");
+                if (e != null) {
+                    val = e.getTextTrim();
+                }
+            }
+            HSSFCell c = row.createCell(i);
+            if (NumberUtils.isNumber(val)) {
+                c.setCellValue(Double.parseDouble(val));
+                c.setCellType(CellType.NUMERIC);
+            } else {
+                c.setCellValue(val);
+            }
+            int rowSpan = NumberUtils.toInt(thEle.attributeValue("rowspan"), 1);
+            int colSpan = NumberUtils.toInt(thEle.attributeValue("colspan"), 1);
+            c.setCellStyle(cellStyle);
+            if (rowSpan > 1 || colSpan > 1) { // 存在跨行或跨列
+                crossRowEleMetaLs.add(new CrossRangeCellMeta(rowIndex, i, rowSpan, colSpan));
+            }
+            if (colSpan > 1) {// 当前行跨列处理（补单元格）
+                for (int j = 1; j < colSpan; j++) {
+                    i++;
+                    row.createCell(i);
+                }
+            }
+        }
+        return i;
+    }
+
+    /**
+     * 设置合并单元格的边框样式
+     *
+     * @param sheet
+     * @param region
+     * @param cs
+     */
+    public static void setRegionStyle(HSSFSheet sheet, CellRangeAddress region, HSSFCellStyle cs) {
+        for (int i = region.getFirstRow(); i <= region.getLastRow(); i++) {
+            HSSFRow row = sheet.getRow(i);
+            for (int j = region.getFirstColumn(); j <= region.getLastColumn(); j++) {
+                HSSFCell cell = row.getCell(j);
+                cell.setCellStyle(cs);
+            }
+        }
+    }
+
+    /**
+     * 获得因rowSpan占据的单元格
+     *
+     * @param rowIndex 行号
+     * @param colIndex 列号
+     * @param crossRowEleMetaLs 跨行列元数据
+     * @return 当前行在某列需要占据单元格
+     */
+    private static int getCaptureCellSize(int rowIndex, int colIndex, List<CrossRangeCellMeta> crossRowEleMetaLs) {
+        int captureCellSize = 0;
+        for (CrossRangeCellMeta crossRangeCellMeta : crossRowEleMetaLs) {
+            if (crossRangeCellMeta.getFirstRow() < rowIndex && crossRangeCellMeta.getLastRow() >= rowIndex) {
+                if (crossRangeCellMeta.getFirstCol() <= colIndex && crossRangeCellMeta.getLastCol() >= colIndex) {
+                    captureCellSize = crossRangeCellMeta.getLastCol() - colIndex + 1;
+                }
+            }
+        }
+        return captureCellSize;
+    }
+
+    /**
+     * 获得标题样式
+     *
+     * @param workbook
+     * @return
+     */
+    private static HSSFCellStyle getTitleStyle(HSSFWorkbook workbook) {
+        //short titlebackgroundcolor = IndexedColors.GREY_25_PERCENT.index;
+        short fontSize = 12;
+        String fontName = "宋体";
+        HSSFCellStyle style = workbook.createCellStyle();
+        style.setVerticalAlignment(VerticalAlignment.CENTER);
+        style.setAlignment(HorizontalAlignment.CENTER);
+        style.setBorderBottom(BorderStyle.THIN); //下边框
+        style.setBorderLeft(BorderStyle.THIN);//左边框
+        style.setBorderTop(BorderStyle.THIN);//上边框
+        style.setBorderRight(BorderStyle.THIN);//右边框
+        //style.setFillPattern(FillPatternType.SOLID_FOREGROUND);
+        //style.setFillForegroundColor(titlebackgroundcolor);// 背景色
+
+        HSSFFont font = workbook.createFont();
+        font.setFontName(fontName);
+        font.setFontHeightInPoints(fontSize);
+        font.setBold(true);
+        style.setFont(font);
+        return style;
+    }
+
+    /**
+     * 获得内容样式
+     *
+     * @param wb
+     * @return
+     */
+    private static HSSFCellStyle getContentStyle(HSSFWorkbook wb) {
+        short fontSize = 12;
+        String fontName = "宋体";
+        HSSFCellStyle style = wb.createCellStyle();
+        style.setBorderBottom(BorderStyle.THIN); //下边框
+        style.setBorderLeft(BorderStyle.THIN);//左边框
+        style.setBorderTop(BorderStyle.THIN);//上边框
+        style.setBorderRight(BorderStyle.THIN);//右边框
+        HSSFFont font = wb.createFont();
+        font.setFontName(fontName);
+        font.setFontHeightInPoints(fontSize);
+        style.setFont(font);
+        style.setAlignment(HorizontalAlignment.CENTER);//水平居中
+        style.setVerticalAlignment(VerticalAlignment.CENTER);//垂直居中
+        style.setWrapText(true);
+        return style;
+    }
+}
+

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/utils/CrossRangeCellMeta.java

@@ -0,0 +1,42 @@
+package cn.smartjavaai.ocr.utils;
+
+/**
+ * @Auther: xiaoqiang
+ * @Date: 2020/12/9 9:17
+ * @Description:
+ */
+public class CrossRangeCellMeta {
+    public CrossRangeCellMeta(int firstRowIndex, int firstColIndex, int rowSpan, int colSpan) {
+        super();
+        this.firstRowIndex = firstRowIndex;
+        this.firstColIndex = firstColIndex;
+        this.rowSpan = rowSpan;
+        this.colSpan = colSpan;
+    }
+
+    private int firstRowIndex;
+    private int firstColIndex;
+    private int rowSpan;// 跨越行数
+    private int colSpan;// 跨越列数
+
+    public int getFirstRow() {
+        return firstRowIndex;
+    }
+
+    public int getLastRow() {
+        return firstRowIndex + rowSpan - 1;
+    }
+
+    public int getFirstCol() {
+        return firstColIndex;
+    }
+
+    public int getLastCol() {
+        return firstColIndex + colSpan - 1;
+    }
+
+    public int getColSpan(){
+        return colSpan;
+    }
+}
+

smartjavaai-ocr/src/main/java/cn/smartjavaai/ocr/utils/OcrUtils.java

@@ -11,14 +11,12 @@ import ai.djl.ndarray.NDManager;
 import ai.djl.opencv.OpenCVImageFactory;
 import cn.smartjavaai.common.entity.*;
 import cn.smartjavaai.common.entity.Point;
-import cn.smartjavaai.ocr.entity.OcrBox;
-import cn.smartjavaai.ocr.entity.OcrInfo;
-import cn.smartjavaai.ocr.entity.OcrItem;
-import cn.smartjavaai.ocr.entity.RotatedBoxCompX;
+import cn.smartjavaai.ocr.entity.*;
 import cn.smartjavaai.ocr.enums.AngleEnum;
 import cn.smartjavaai.ocr.opencv.OcrNDArrayUtils;
 import cn.smartjavaai.ocr.opencv.OcrOpenCVUtils;
 import lombok.extern.slf4j.Slf4j;
+import org.apache.commons.collections.CollectionUtils;
 import org.opencv.core.Mat;
 import org.opencv.core.Scalar;
 import org.opencv.imgproc.Imgproc;
@@ -26,10 +24,8 @@ import org.opencv.imgproc.Imgproc;
 import java.awt.*;
 import java.awt.image.BufferedImage;
 import java.math.BigDecimal;
-import java.util.ArrayList;
-import java.util.Iterator;
+import java.util.*;
 import java.util.List;
-import java.util.Objects;
 
 /**
  * @author dwj
@@ -42,27 +38,41 @@ public class OcrUtils {
     /**
      * 转换为OcrBox
      * @param dt_boxes
-     * @param img
      * @return
      */
-    public static List<OcrBox> convertToOcrBox(NDList dt_boxes, Image img){
-        if(Objects.isNull(dt_boxes) || dt_boxes.size() == 0){
-            return null;
-        }
-        List<OcrBox> boxList = new ArrayList<OcrBox>();
-        for(NDArray box : dt_boxes){
+    public static List<OcrBox> convertToOcrBox(NDList dt_boxes) {
+        List<OcrBox> boxList = new ArrayList<>();
+        for (NDArray box : dt_boxes) {
             float[] pointsArr = box.toFloatArray();
-            //log.debug("points: {}", pointsArr);
-            float[] lt = java.util.Arrays.copyOfRange(pointsArr, 0, 2);
-            float[] rt = java.util.Arrays.copyOfRange(pointsArr, 2, 4);
-            float[] rb = java.util.Arrays.copyOfRange(pointsArr, 4, 6);
-            float[] lb = java.util.Arrays.copyOfRange(pointsArr, 6, 8);
-            OcrBox ocrBox = new OcrBox(new Point(lt[0], lt[1]), new Point(rt[0], rt[1]), new Point(rb[0], rb[1]), new Point(lb[0], lb[1]));
+            OcrBox ocrBox = new OcrBox(
+                    new Point(pointsArr[0], pointsArr[1]),
+                    new Point(pointsArr[2], pointsArr[3]),
+                    new Point(pointsArr[4], pointsArr[5]),
+                    new Point(pointsArr[6], pointsArr[7])
+            );
             boxList.add(ocrBox);
         }
         return boxList;
     }
 
+    /**
+     * 转换为OcrBox
+     * @param dt_boxes
+     * @return
+     */
+    public static List<List<OcrBox>> convertToOcrBox(List<NDList> ndLists) {
+        if (ndLists == null || ndLists.isEmpty()) {
+            return Collections.emptyList();
+        }
+        List<List<OcrBox>> boxLists = new ArrayList<>();
+        for (NDList dt_boxes : ndLists) {
+            boxLists.add(convertToOcrBox(dt_boxes));
+        }
+        return boxLists;
+    }
+
+
+
     /**
      * 欧式距离计算
      *
@@ -140,7 +150,6 @@ public class OcrUtils {
         if(Objects.isNull(lines) || lines.size() == 0){
             return null;
         }
-        List<DetectionInfo> detectionInfoList = new ArrayList<DetectionInfo>();
         List<List<OcrItem>> lineList = new ArrayList<List<OcrItem>>();
         String fullText = "";
         for(ArrayList<RotatedBoxCompX> boxList : lines){
@@ -165,6 +174,36 @@ public class OcrUtils {
         return new OcrInfo(lineList, fullText);
     }
 
+    public static OcrInfo convertRotatedBoxesToOcrItems(List<RotatedBox> rotatedBoxes) {
+        OcrInfo ocrInfo = new OcrInfo();
+        List<OcrItem> ocrItems = new ArrayList<>();
+        StringBuilder fullText = new StringBuilder();
+        for (RotatedBox rotatedBox : rotatedBoxes) {
+            NDArray box = rotatedBox.getBox();
+            float[] points = box.toFloatArray();
+            Point topLeft = new Point(points[0], points[1]);
+            Point topRight = new Point(points[2], points[3]);
+            Point bottomRight = new Point(points[4], points[5]);
+            Point bottomLeft = new Point(points[6], points[7]);
+
+            OcrBox ocrBox = new OcrBox(topLeft, topRight, bottomRight, bottomLeft);
+            String text = rotatedBox.getText();
+
+            OcrItem item = new OcrItem();
+            item.setOcrBox(ocrBox);
+            item.setText(text);
+            ocrItems.add(item);
+            fullText.append(text + " ");
+        }
+        if (fullText.length() > 0) {
+            fullText.deleteCharAt(fullText.length() - 1);
+        }
+        ocrInfo.setOcrItemList(ocrItems);
+        ocrInfo.setFullText(fullText.toString());
+        return ocrInfo;
+    }
+
+
 
     /**
      * 放射变换+裁剪
@@ -235,26 +274,28 @@ public class OcrUtils {
             // 声明画笔属性 ：粗 细（单位像素）末端无修饰 折线处呈尖角
             BasicStroke bStroke = new BasicStroke(2, BasicStroke.CAP_BUTT, BasicStroke.JOIN_MITER);
             g.setStroke(bStroke);
-            for(List<OcrItem> ocrItemList : ocrInfo.getLineList()){
-                for(OcrItem item : ocrItemList){
-                    OcrBox box = item.getOcrBox();
-                    int[] xPoints = {
-                            (int)box.getTopLeft().getX(),
-                            (int)box.getTopRight().getX(),
-                            (int)box.getBottomRight().getX(),
-                            (int)box.getBottomLeft().getX(),
-                            (int)box.getTopLeft().getX()
-                    };
-                    int[] yPoints = {
-                            (int)box.getTopLeft().getY(),
-                            (int)box.getTopRight().getY(),
-                            (int)box.getBottomRight().getY(),
-                            (int)box.getBottomLeft().getY(),
-                            (int)box.getTopLeft().getY()
-                    };
-                    g.drawPolyline(xPoints, yPoints, 5);
-                    g.drawString(item.getText(), xPoints[0], yPoints[0]);
-                }
+            List<OcrItem> ocrItemList = ocrInfo.getOcrItemList();
+            if(CollectionUtils.isNotEmpty(ocrInfo.getLineList())){
+                ocrItemList = ocrInfo.flattenLines();
+            }
+            for(OcrItem item : ocrItemList){
+                OcrBox box = item.getOcrBox();
+                int[] xPoints = {
+                        (int)box.getTopLeft().getX(),
+                        (int)box.getTopRight().getX(),
+                        (int)box.getBottomRight().getX(),
+                        (int)box.getBottomLeft().getX(),
+                        (int)box.getTopLeft().getX()
+                };
+                int[] yPoints = {
+                        (int)box.getTopLeft().getY(),
+                        (int)box.getTopRight().getY(),
+                        (int)box.getBottomRight().getY(),
+                        (int)box.getBottomLeft().getY(),
+                        (int)box.getTopLeft().getY()
+                };
+                g.drawPolyline(xPoints, yPoints, 5);
+                g.drawString(item.getText(), xPoints[0], yPoints[0]);
             }
         } finally {
             g.dispose();