fix(exam): 优化考试单词识别与文件上传功能

- 修改application.yml默认激活环境为pro，调整上传文件大小限制为30MB - 更新pro配置中的模板文件版本和路径 - 修复考试单词ID乱序问题，添加生成试卷成功日志 - 优化考试识别逻辑，确保ansSheetPath变量正确处理，完善异常捕获与文件删除机制 - 增加上传文件格式白名单校验，强化文件存储路径和命名安全性 - 用NIO替换File操作，确保上传目录存在并合理创建 - 优化PngUtil图像二值化处理，增加自适应阈值和形态学操作减少噪声 - 修改未背熟单词识别阈值，调整检测区域坐标和日志输出 - 注释部分冗余图像预处理代码，完善日志和异常信息提示 - 统一文件上传与识别过程的错误处理和日志记录，提升系统稳定性和可维护性
2026-01-06 14:48:01 +08:00
7 changed files with 141 additions and 85 deletions
--- a/enlish-service/src/main/java/com/yinlihupo/enlish/service/controller/ExamWordsController.java
+++ b/enlish-service/src/main/java/com/yinlihupo/enlish/service/controller/ExamWordsController.java
@@ -62,7 +62,7 @@ public class ExamWordsController {
            // bug: 获取单词后，单词的id会乱序、 需要重新更新考试记录中的 id
            examWordsDO.setWordIds(assessmentWords.stream().map(Word::getId).toList());
            examWordsService.updateExamWordsWordIdsOrder(examWordsDO);
-
+            log.info("生成试卷成功 {}", examWordsDO);
            List<StudentDetail> studentDetailList = studentService.getStudentDetailList(Collections.singletonList(studentId));
            List<Map<String, Object>> maps = studentDetailList.stream().map(studentDetail -> {
                Map<String, Object> data = new HashMap<>();
--- a/enlish-service/src/main/java/com/yinlihupo/enlish/service/service/exam/ExamWordsServiceImpl.java
+++ b/enlish-service/src/main/java/com/yinlihupo/enlish/service/service/exam/ExamWordsServiceImpl.java
@@ -10,16 +10,17 @@ import org.checkerframework.checker.nullness.qual.NonNull;
 import org.springframework.beans.factory.annotation.Value;
 import org.springframework.stereotype.Service;
 import org.springframework.transaction.annotation.Transactional;
+import org.springframework.util.StringUtils;
 import org.springframework.web.multipart.MultipartFile;

 import java.io.File;
 import java.io.IOException;
+import java.nio.file.Files;
+import java.nio.file.Path;
+import java.nio.file.Paths;
 import java.time.LocalDate;
 import java.time.LocalDateTime;
-import java.util.ArrayList;
-import java.util.Collections;
-import java.util.List;
-import java.util.UUID;
+import java.util.*;

@Service
@Slf4j
@@ -118,7 +119,7 @@ public class ExamWordsServiceImpl implements ExamWordsService {
        StudentDO studentDO = studentDOMapper.selectStudentById(studentId);
        Integer gradeId = studentDO.getGradeId();
        List<UnitDO> unitDOS = unitDOMapper.selectByUnitName(ExamWordsConstant.getGradeName(gradeId) + "上");
-        ExamWordsDO examWordsDO = getExamWordsDO(studentId, studentDO, gradeId, unitDOS);
+        ExamWordsDO examWordsDO = getExamWordsDO(studentId, studentDO, gradeId, unitDOS, ExamWordsConstant.EXAM_TYPE_MIDTERM);
        examWordsDO.setTitle("期中测试" + studentDO.getName());
        return getExamWordsDO(studentId, examWordsDO);

@@ -128,13 +129,13 @@ public class ExamWordsServiceImpl implements ExamWordsService {
        StudentDO studentDO = studentDOMapper.selectStudentById(studentId);
        Integer gradeId = studentDO.getGradeId();
        List<UnitDO> unitDOS = unitDOMapper.selectByUnitName(ExamWordsConstant.getGradeName(gradeId));
-        ExamWordsDO examWordsDO = getExamWordsDO(studentId, studentDO, gradeId, unitDOS);
+        ExamWordsDO examWordsDO = getExamWordsDO(studentId, studentDO, gradeId, unitDOS, ExamWordsConstant.EXAM_TYPE_FINAL);
        examWordsDO.setTitle("期末测试" + studentDO.getName());
        return getExamWordsDO(studentId, examWordsDO);
    }

    @NonNull
-    private ExamWordsDO getExamWordsDO(Integer studentId, StudentDO studentDO, Integer gradeId, List<UnitDO> unitDOS) {
+    private ExamWordsDO getExamWordsDO(Integer studentId, StudentDO studentDO, Integer gradeId, List<UnitDO> unitDOS, Integer type) {
        if (unitDOS.isEmpty()) {
            throw new RuntimeException("没有找到对应的单元");
        }
@@ -143,7 +144,7 @@ public class ExamWordsServiceImpl implements ExamWordsService {
        ExamWordsDO examWordsDO = ExamWordsDO.builder()
                .gradeId(gradeId)
                .level(1)
-                .type(ExamWordsConstant.EXAM_TYPE_BASELINE)
+                .type(type)
                .title(studentDO.getName())
                .createdAt(LocalDateTime.now())
                .wordIds(vocabularyBankDOS.stream().map(VocabularyBankDO::getId).toList())
@@ -170,30 +171,55 @@ public class ExamWordsServiceImpl implements ExamWordsService {
    @Override
    @Transactional(rollbackFor = RuntimeException.class)
    public int saveExamWordsPngToDbAndLocal(MultipartFile file) {
-
-        File dir = new File(tmpPng);
-        if (!dir.exists()) {
-            dir.mkdirs();
+        // 1. 基础校验：判空
+        if (file == null || file.isEmpty()) {
+            throw new RuntimeException("上传文件不能为空");
        }

-        try {
+        // 2. 安全校验：检查后缀名白名单
        String originalFilename = file.getOriginalFilename();
-            String suffix = "";
-            if (originalFilename != null && originalFilename.contains(".")) {
-                suffix = originalFilename.substring(originalFilename.lastIndexOf("."));
-            }
-            String newFileName = UUID.randomUUID() + suffix;
-            String path = tmpPng + newFileName;
+        String extension = StringUtils.getFilenameExtension(originalFilename); // Spring工具类
+        List<String> allowedExtensions = Arrays.asList("png", "jpg", "jpeg");

-            File dest = new File(path);
-            file.transferTo(dest);
-            int insert = examWordsJudgeResultDOMapper.insert(path);
-            log.info("上传文件成功");
+        if (extension == null || !allowedExtensions.contains(extension.toLowerCase())) {
+            throw new RuntimeException("不支持的文件格式，仅支持: " + allowedExtensions);
+        }
+
+        // 3. 准备目录 (使用 NIO)
+        // 假设 tmpPng 是配置好的基础路径字符串
+        Path directoryPath = Paths.get(tmpPng);
+        try {
+            if (!Files.exists(directoryPath)) {
+                Files.createDirectories(directoryPath);
+            }
+
+            // 4. 生成文件名 (防止文件名冲突)
+            String newFileName = UUID.randomUUID().toString() + "." + extension;
+
+            // 5. 组合最终路径 (自动处理分隔符)
+            Path targetPath = directoryPath.resolve(newFileName);
+
+            // 6. 保存文件
+            file.transferTo(targetPath.toAbsolutePath().toFile());
+            String string = targetPath.toAbsolutePath().toFile().toString();
+            log.info("文件上传成功路径为 {}", string);
+            if (!targetPath.toFile().exists()) {
+                log.error("文件上传失败: {}", newFileName);
+                throw new RuntimeException("文件上传失败");
+            }
+
+            // 7. 入库
+            // 建议：存相对路径或文件名，不要存 targetPath.toString() 的绝对路径
+            // 这里为了演示，假设 insert 依然接受字符串，建议存 newFileName
+            int insert = examWordsJudgeResultDOMapper.insert(targetPath.toString());
+
+            log.info("上传文件成功: {}", newFileName);
            return insert;
-        } catch (IOException e) {
-            throw new RuntimeException("上传失败", e);
-        }

+        } catch (IOException e) {
+            log.error("文件上传失败: {}", originalFilename, e);
+            throw new RuntimeException("上传失败，请稍后重试", e);
+        }
    }

    @Override
--- a/enlish-service/src/main/java/com/yinlihupo/enlish/service/service/judge/ExamWordsJudgeServiceImpl.java
+++ b/enlish-service/src/main/java/com/yinlihupo/enlish/service/service/judge/ExamWordsJudgeServiceImpl.java
@@ -53,9 +53,11 @@ public class ExamWordsJudgeServiceImpl implements ExamWordsJudgeService {
    public void judgeExamWords(int count) {
        List<ExamWordsJudgeResultDO> examWordsJudgeResultDOS = examWordsJudgeResultDOMapper.selectUnfinishedExamWordsJudgeResultDOList(count);
        for (ExamWordsJudgeResultDO examWordsJudgeResultDO : examWordsJudgeResultDOS) {
+            String ansSheetPath = null;
            try {
-                String ansSheetPath = examWordsJudgeResultDO.getAnsSheetPath();
+                ansSheetPath = examWordsJudgeResultDO.getAnsSheetPath();
                List<CoordinatesXY> coordinatesXIES = PngUtil.analysisXY(ansSheetPath);
+
                // 从图片中获取学生 id 和考试 id
                StudentExamId studentExamId = PngUtil.analyzeExamWordsIdAndStudentId(ansSheetPath, tessdataPath, coordinatesXIES);
                Integer examWordsJudgeResultDOId = examWordsJudgeResultDO.getId();
@@ -79,7 +81,7 @@ public class ExamWordsJudgeServiceImpl implements ExamWordsJudgeService {
                }

                ExamWordsDO examWordsDO = examWordsDOMapper.selectById(examWordsId);
-                if(examWordsDO == null) {
+                if (examWordsDO == null) {
                    examWordsJudgeResultDOMapper.updateMsg(examWordsJudgeResultDOId, "未找到考试");
                    continue;
                }
@@ -146,9 +148,14 @@ public class ExamWordsJudgeServiceImpl implements ExamWordsJudgeService {
                boolean delete = new File(ansSheetPath).delete();
                if (delete) {
                    log.info("删除文件成功：{}", ansSheetPath);
+                } else {
+                    log.error("删除文件失败：{}", ansSheetPath);
                }
            } catch (Exception e) {
                log.error("识别考试失败 {}", e.getMessage());
+                if (ansSheetPath != null) {
+                    new File(ansSheetPath).delete();
+                }
                examWordsJudgeResultDOMapper.updateMsg(examWordsJudgeResultDO.getId(), e.getMessage());
            }
        }
--- a/enlish-service/src/main/java/com/yinlihupo/enlish/service/utils/PngUtil.java
+++ b/enlish-service/src/main/java/com/yinlihupo/enlish/service/utils/PngUtil.java
@@ -15,6 +15,7 @@ import org.opencv.imgproc.Imgproc;

 import java.awt.image.BufferedImage;
 import java.awt.image.DataBufferByte;
+import java.io.File;
 import java.util.ArrayList;
 import java.util.Comparator;
 import java.util.List;
@@ -31,31 +32,13 @@ public class PngUtil {
    // 获取起始坐标
    public static List<CoordinatesXY> analysisXY(String imagePath) {

+        Mat binary = image2BinaryMath(imagePath);
        Mat src = Imgcodecs.imread(imagePath);
-
-        if (src.empty()) {
-            System.out.println("无法读取图片，请检查路径。");
-            return null;
-        }
-
-        // 3. 预处理
-        // 3.1 转换为灰度图
-        Mat gray = new Mat();
-        Imgproc.cvtColor(src, gray, Imgproc.COLOR_BGR2GRAY);
-
-        // 3.2 二值化处理 (Thresholding)
-        // 使用 THRESH_BINARY_INV (反转二值化)，因为我们需要找的是白色背景上的黑色块。
-        // 反转后，黑色块变成白色(255)，背景变成黑色(0)，方便 findContours 查找。
-        Mat binary = new Mat();
-        // 阈值设为 50 左右即可，因为块是纯黑的
-        Imgproc.threshold(gray, binary, 80, 255, Imgproc.THRESH_BINARY_INV);
-
        // 4. 查找轮廓
        List<MatOfPoint> contours = new ArrayList<>();
        Mat hierarchy = new Mat();
-        // RETR_EXTERNAL 只检测最外层轮廓，忽略块内部可能存在的噪点
        Imgproc.findContours(binary, contours, hierarchy, Imgproc.RETR_EXTERNAL, Imgproc.CHAIN_APPROX_SIMPLE);
-        //Imgcodecs.imwrite("output_red___v1.png", binary);
+
        System.out.println("检测到的轮廓总数: " + contours.size());
        System.out.println("------------------------------------------------");

@@ -91,36 +74,31 @@ public class PngUtil {
                // 可选：在原图上画出框，用于调试验证
                Imgproc.rectangle(src, rect, new Scalar(0, 0, 255), 2); // 红色框
                Imgproc.putText(src, "#" + blockCount, new Point(rect.x, rect.y - 5), Imgproc.FONT_HERSHEY_SIMPLEX, 0.5, new Scalar(0, 0, 255), 1);
-                //Imgcodecs.imwrite("output_red.png", src);
            }
        }
+        Imgcodecs.imwrite("output_red.png", src);
        System.out.println("找到 " + blockCount + " 个黑色块。");

-
        // 计算起始坐标
        list.sort(Comparator.comparingInt(CoordinatesXY::getX));

+        list.forEach(coordinatesXY -> coordinatesXY.setHeight(coordinatesXY.getHeight() / 51));
+        list.forEach(coordinatesXY -> coordinatesXY.setWidth(coordinatesXY.getWidth() / 3));
+        list.forEach(coordinatesXY -> coordinatesXY.setX(coordinatesXY.getX() + coordinatesXY.getWidth() * 2));
+
+        log.info("起始坐标: {}", list);
+
        return list;
    }

    // 获取（未背熟）单词的 id
    public static List<Integer> analyzePngForUnmemorizedWordIds(String filePath, List<Integer> wordIds, List<CoordinatesXY> coordinatesXYList) {

-        Mat src = Imgcodecs.imread(filePath);
-        if (src.empty()) {
-            log.error("无法读取图片，请检查路径: {}", filePath);
-            throw new RuntimeException("无法读取图片");
-        }

-        Mat gray = new Mat();
-        Mat binary = new Mat();
+        Mat binary = image2BinaryMath(filePath);

        try {
-            Imgproc.cvtColor(src, gray, Imgproc.COLOR_BGR2GRAY);
-            // 建议：如果光照不均匀，考虑使用 THRESH_OTSU 自动阈值，或者自适应阈值
-            Imgproc.threshold(gray, binary, 150, 255, Imgproc.THRESH_BINARY_INV);
-//            调试时打印
-            //Imgcodecs.imwrite("output_binary.png", binary);
+
            List<Integer> answer = new ArrayList<>();
            int words_index = 0;

@@ -150,8 +128,8 @@ public class PngUtil {
                    Rect rect = new Rect(currentX + 1, currentY + 1, width - 2, height - 2);
                    Mat region = binary.submat(rect);
                    int countNonZero = Core.countNonZero(region);
-
-                    if (countNonZero > 370) {
+                    log.info("当前位置为 words_index={}，坐标为 x={} y={} 当前区域非零像素数: {}", words_index, currentX, currentY, countNonZero);
+                    if (countNonZero > 1000) {
                        Integer id = wordIds.get(words_index);
                        answer.add(id);
                        log.info("检测到标记（未背熟）：ID={}, 当前坐标 x = {} y = {} ", id, currentX + 1, currentY + 1);
@@ -170,8 +148,6 @@ public class PngUtil {

        } finally {

-            src.release();
-            gray.release();
            binary.release();
        }
    }
@@ -191,21 +167,21 @@ public class PngUtil {
        Rect roiRect = new Rect(0, 0, left.getX(), left.getY());
        Mat roi = new Mat(src, roiRect);

-        // 3. 图像预处理 (提高 OCR 准确率)
-        // 3.1 转为灰度图
-        Mat gray = new Mat();
-        Imgproc.cvtColor(roi, gray, Imgproc.COLOR_BGR2GRAY);
-
-        // 3.2 二值化 (Thresholding)
-        // 使用 OTSU 算法自动寻找最佳阈值，或者手动指定阈值
-        Mat binary = new Mat();
-        Imgproc.threshold(gray, binary, 0, 255, Imgproc.THRESH_BINARY | Imgproc.THRESH_OTSU);
+//        // 3. 图像预处理 (提高 OCR 准确率)
+//        // 3.1 转为灰度图
+//        Mat gray = new Mat();
+//        Imgproc.cvtColor(roi, gray, Imgproc.COLOR_BGR2GRAY);
+//
+//        // 3.2 二值化 (Thresholding)
+//        // 使用 OTSU 算法自动寻找最佳阈值，或者手动指定阈值
+//        Mat binary = new Mat();
+//        Imgproc.threshold(gray, binary, 0, 255, Imgproc.THRESH_BINARY | Imgproc.THRESH_OTSU);

        // 可选：保存预处理后的图片查看效果
-        //Imgcodecs.imwrite("debug_roi.jpg", binary);
+        Imgcodecs.imwrite("debug_roi.jpg", src);

        // 4. 将 OpenCV Mat 转换为 BufferedImage (供 Tess4J 使用)
-        BufferedImage processedImage = matToBufferedImage(binary);
+        BufferedImage processedImage = matToBufferedImage(src);

        // 5. 使用 Tesseract 进行 OCR 识别
        ITesseract instance = new Tesseract();
@@ -230,6 +206,50 @@ public class PngUtil {
        return null;
    }

+    private static Mat image2BinaryMath(String imagePath) {
+
+        if (!new File(imagePath).exists()) {
+            log.error("图片不存在，请检查路径: {}", imagePath);
+            throw new RuntimeException("图片不存在");
+        }
+
+        Mat src = Imgcodecs.imread(imagePath);
+
+        if (src.empty()) {
+            log.info("无法读取图片，请检查路径: {}", imagePath);
+            throw new RuntimeException("无法读取图片");
+        }
+
+        Mat gray = new Mat();
+        //转换为灰度图
+        Imgproc.cvtColor(src, gray, Imgproc.COLOR_BGR2GRAY);
+
+        Imgproc.GaussianBlur(gray, gray, new Size(5, 5), 0);
+
+        Mat binary = new Mat();
+        Imgproc.adaptiveThreshold(gray, binary, 255,
+                Imgproc.ADAPTIVE_THRESH_GAUSSIAN_C,
+                Imgproc.THRESH_BINARY_INV, 25, 10);
+
+        Mat kernel = Imgproc.getStructuringElement(Imgproc.MORPH_RECT, new Size(3, 3));
+
+        //开运算 (Open)：先腐蚀后膨胀，用于去除背景中的微小噪点
+        Imgproc.morphologyEx(binary, binary, Imgproc.MORPH_OPEN, kernel);
+
+        // 闭运算 (Close)：先膨胀后腐蚀，用于连接断裂的区域并填充块内部的空洞
+        // 如果块比较大且内部反光严重，可以将 Size(3,3) 改为 Size(5,5) 或更大
+        Mat closeKernel = Imgproc.getStructuringElement(Imgproc.MORPH_RECT, new Size(7, 7));
+        Imgproc.morphologyEx(binary, binary, Imgproc.MORPH_CLOSE, closeKernel);
+
+        // 保存二值化过程图用于调试 (生产环境可注释)
+        Imgcodecs.imwrite("debug_binary_natural.png", binary);
+
+        src.release();
+        gray.release();
+
+        return binary;
+    }
+
    private static @NonNull StudentExamId getStudentExamId(Pattern pattern, String result) {
        Matcher matcher = pattern.matcher(result);
        StudentExamId studentExamId = new StudentExamId(0, 0);
--- a/enlish-service/src/main/resources/config/application-dev.yml
+++ b/enlish-service/src/main/resources/config/application-dev.yml
@@ -39,7 +39,7 @@ templates:
        weekend: C:\project\java\enlish_edu\enlish\enlish-service\src\main\resources\templates\study_plan_review_v3.docx
    plan_day: 7
 tmp:
-    png: C:\project\java\enlish_edu\enlish\enlish-service\src\main\resources\tmp\png\
+    png:

 ai:
    key: app-loC6IrJpj4cS54MAYp73QtGl
--- a/enlish-service/src/main/resources/config/application-pro.yml
+++ b/enlish-service/src/main/resources/config/application-pro.yml
@@ -31,15 +31,15 @@ spring:


 templates:
-    word: assessment_v7.docx
+    word: assessment_v9.docx
    count: 100
-    data: eng.traineddata
+    data:
    plan:
-        weekday: tem_study_plan_v6.docx
-        weekend: study_plan_review_v2.docx
+        weekday: tem_study_plan_v7.docx
+        weekend: study_plan_review_v3.docx
    plan_day: 7
 tmp:
-    png: tmp\png\
+    png:

 ai:
    key: app-loC6IrJpj4cS54MAYp73QtGl
--- a/enlish-service/src/main/resources/config/application.yml
+++ b/enlish-service/src/main/resources/config/application.yml
@@ -3,8 +3,11 @@ server:

 spring:
    profiles:
-        active: dev # 默认激活 dev 本地开发环境
-
+        active: pro # 默认激活 dev 本地开发环境
+    servlet:
+        multipart:
+            max-file-size: 30MB
+            max-request-size: 30MB
 mybatis:
    # MyBatis xml 配置文件路径
    mapper-locations: classpath:/mapper/**/*.xml