diff --git a/src/main/java/com/example/lxcameraapi/controller/CategoryController.java b/src/main/java/com/example/lxcameraapi/controller/CategoryController.java
index dab8b49..ab1a89c 100644
--- a/src/main/java/com/example/lxcameraapi/controller/CategoryController.java
+++ b/src/main/java/com/example/lxcameraapi/controller/CategoryController.java
@@ -55,7 +55,7 @@ public class CategoryController {
 
 
             String path = appConfig.getPicPath() + formattedDate + "/" + request.getCameraId() + "/" + formattedTime + ".png";
-            String url = appConfig.getPicUrl() + formattedDate + "/" + request.getCameraId() + "/" + formattedTime + ".png";
+            String url = appConfig.getPicUrl() + formattedDate + "/" + request.getCameraId() + "/" + formattedTime + ".png"+"detect.jpg";
             log.info("开始拍照，IP: {}, 路径: {}", request.getCameraId(), path);
             String ip = null;
             for (AppConfig.Camera camera : appConfig.getHikCamera()) {
diff --git a/src/main/java/com/example/lxcameraapi/controller/QtCodeController.java b/src/main/java/com/example/lxcameraapi/controller/QtCodeController.java
index 58cb11c..6dda3a6 100644
--- a/src/main/java/com/example/lxcameraapi/controller/QtCodeController.java
+++ b/src/main/java/com/example/lxcameraapi/controller/QtCodeController.java
@@ -283,7 +283,7 @@ public class QtCodeController {
         boxCountResponse.setSuccess(true);
         boxCountResponse.setResult("Processing");
 
-        boxCountResponse.setImagePath(imageUrls.get(0));
+        boxCountResponse.setImagePath(imageUrls.get(0)+ ".jpg");
         boxCountResponse.setTaskId(request.getTaskId());
         boxCountResponse.setScanType(request.getScanType());
         boxCountResponse.setQrCodeResults(imageUrls);
diff --git a/src/main/java/com/example/lxcameraapi/service/AsyncProcessingService.java b/src/main/java/com/example/lxcameraapi/service/AsyncProcessingService.java
index a6fd13f..4d8bb02 100644
--- a/src/main/java/com/example/lxcameraapi/service/AsyncProcessingService.java
+++ b/src/main/java/com/example/lxcameraapi/service/AsyncProcessingService.java
@@ -213,7 +213,10 @@ public class AsyncProcessingService {
 
     /**
      * 异步处理多图片二维码识别
-     * 所有图片识别完成后统一发送结果
+     * 优化流程：
+     * 1. 所有图片先做 ONNX 检测，收集全部候选框
+     * 2. 合并后 NMS 过滤近距框
+     * 3. 每个框依次尝试不同图片的裁剪解码，一张成功即跳过
      */
     @Async
     public void processMultiQrCodeAsync(Integer cameraId, String taskId, List<String> imagePaths,
@@ -222,71 +225,107 @@ public class AsyncProcessingService {
         try {
             log.info("异步多图片识别任务开始，图片数量: {}", imagePaths.size());
 
-            // 收集所有图片的二维码结果
-            List<String> allQrCodeResults = new ArrayList<>();
-            List<String> allImagePaths = new ArrayList<>();
+            // ====== Phase 1: 批量检测，收集所有候选框 ======
+            List<BoundingBox> allDetections = new ArrayList<>();
+            // 缓存所有图片的 BufferedImage，避免重复读取
+            List<BufferedImage> cachedImages = new ArrayList<>();
 
             for (int i = 0; i < imagePaths.size(); i++) {
                 String path = imagePaths.get(i);
-                String url = imageUrls.get(i);
-
                 try {
-                    log.info("开始识别第{}/{}张图片: {}", i + 1, imagePaths.size(), path);
-
+                    log.info("Phase1 - 检测第{}/{}张图片: {}", i + 1, imagePaths.size(), path);
                     List<BoundingBox> detectResults = onnxServiceNew.detect26(path, "qrCode");
-                    log.info("第{}张图片检测结果: {}", i + 1, detectResults);
-
-                    // 读取原始图片
-                    BufferedImage originalImage = ImageIO.read(new File(path));
-                    if (originalImage == null) {
-                        log.warn("无法读取图片: {}", path);
-                        continue;
+                    log.info("检测结果: {} 个框", detectResults.size());
+                    allDetections.addAll(detectResults);
+
+                    // 预加载图片到内存
+                    BufferedImage image = ImageIO.read(new File(path));
+                    if (image != null) {
+                        cachedImages.add(image);
+                    } else {
+                        log.warn("无法读取图片: {}, 用null占位", path);
+                        cachedImages.add(null);
                     }
+                } catch (Exception e) {
+                    log.error("第{}张图片检测或加载异常: {}", i + 1, path, e);
+                    cachedImages.add(null);
+                }
+            }
+
+            log.info("Phase1 完成，共检测到 {} 个候选框", allDetections.size());
+
+            if (allDetections.isEmpty()) {
+                log.warn("所有图片均未检测到二维码候选框，发送 Unknown 结果");
+                sendQrCodeResult(targetIp, targetPort, targetPath, taskId, "Unknown",
+                        imageUrls.get(0) + ".jpg", new ArrayList<>(), scanType);
+                return;
+            }
 
-                    List<String> qrCodeResults = new ArrayList<>();
-                    List<BoundingBox> validBoxes = new ArrayList<>();
-
-                    for (int j = 0; j < detectResults.size(); j++) {
-                        BoundingBox box = detectResults.get(j);
-                        try {
-                            BufferedImage croppedImage = cropBoundingBox(originalImage, box, 30);
-                            
-                            // 保存裁剪的原始图片用于调试
-                            String debugPath = path + "_crop_" + i + "_" + j + ".jpg";
-                            ImageIO.write(croppedImage, "jpg", new File(debugPath));
-                            log.info("DEBUG: 裁剪图片保存到: {}, 尺寸: {}x{}, 置信度: {}", 
-                                debugPath, croppedImage.getWidth(), croppedImage.getHeight(), box.getConfidence());
-
-                            // 最小尺寸放大
-                            int minSize = 320;
-                            if (croppedImage.getWidth() < minSize || croppedImage.getHeight() < minSize) {
-                                int newWidth = Math.max(croppedImage.getWidth(), minSize);
-                                int newHeight = Math.max(croppedImage.getHeight(), minSize);
-                                croppedImage = resizeImage(croppedImage, newWidth, newHeight);
-                                log.info("DEBUG: 图片已放大到: {}x{}", newWidth, newHeight);
-                            }
-
-                            // 解码
-                            String qrCodeContent = decodeQrCodeWithRetry(croppedImage, box.getConfidence(), 2);
-
-                            if (qrCodeContent != null) {
-                                qrCodeResults.add(qrCodeContent);
-                                validBoxes.add(box);
-                            }
-                        } catch (Exception e) {
-                            log.error("第{}张图片解码异常，box: {}", i + 1, box, e);
+            // ====== Phase 2: NMS 过滤重叠框 ======
+            List<BoundingBox> filteredBoxes = ONNXServiceNew.nonMaxSuppression(allDetections, 0.5f);
+            log.info("Phase2 NMS完成: {} -> {} 个框", allDetections.size(), filteredBoxes.size());
+
+            // ====== Phase 3: 逐框跨图解码 ======
+            List<String> allQrCodeResults = new ArrayList<>();
+            List<BoundingBox> validBoxes = new ArrayList<>();
+            // 用第一张图片的 url 作为标注图
+            String annotatedImageUrl = imageUrls.get(0) + ".jpg";
+
+            for (int boxIdx = 0; boxIdx < filteredBoxes.size(); boxIdx++) {
+                BoundingBox box = filteredBoxes.get(boxIdx);
+                boolean decoded = false;
+
+                // 遍历每张图片尝试解码当前框
+                for (int imgIdx = 0; imgIdx < cachedImages.size(); imgIdx++) {
+                    BufferedImage image = cachedImages.get(imgIdx);
+                    if (image == null) continue;
+
+                    try {
+                        BufferedImage cropped = cropBoundingBox(image, box, 30);
+
+                        // 保存裁剪图用于调试（仅首张成功前）
+                        if (!decoded) {
+                            String debugPath = imagePaths.get(imgIdx) + "_crop_box" + boxIdx + ".jpg";
+                            ImageIO.write(cropped, "jpg", new File(debugPath));
+                            log.debug("裁剪图保存: {} (来源: 图{})", debugPath, imgIdx + 1);
                         }
-                    }
 
-                    // 绘制检测框
-                    opencvService.drawBoundingBoxesOnImage(validBoxes, path, path + ".jpg");
+                        // 最小尺寸放大
+                        int minSize = 320;
+                        if (cropped.getWidth() < minSize || cropped.getHeight() < minSize) {
+                            int newWidth = Math.max(cropped.getWidth(), minSize);
+                            int newHeight = Math.max(cropped.getHeight(), minSize);
+                            cropped = resizeImage(cropped, newWidth, newHeight);
+                        }
+
+                        String qrCodeContent = decodeQrCodeWithRetry(cropped, box.getConfidence(), 2);
 
-                    // 收集结果
-                    allQrCodeResults.addAll(qrCodeResults);
-                    allImagePaths.add(url + ".jpg");
+                        if (qrCodeContent != null) {
+                            log.info("框{} 解码成功 (来源图{}): {}", boxIdx, imgIdx + 1, qrCodeContent);
+                            allQrCodeResults.add(qrCodeContent);
+                            validBoxes.add(box);
+                            box.setName(qrCodeContent);
+                            decoded = true;
+                            break; // 本框成功，试下一个框
+                        }
+                    } catch (Exception e) {
+                        log.warn("框{} 图{} 解码异常: {}", boxIdx, imgIdx + 1, e.getMessage());
+                    }
+                }
+
+                if (!decoded) {
+                    log.warn("框{} 所有图片均解码失败, 坐标: ({}, {}), 尺寸: {}x{}, 置信度: {}",
+                            boxIdx, (int) box.getX(), (int) box.getY(), (int) box.getW(), (int) box.getH(), box.getConfidence());
+                }
+            }
 
+            // 在第一张图片上绘制所有检测框
+            if (!cachedImages.isEmpty() && cachedImages.get(0) != null) {
+                try {
+                    String annotatePath = imagePaths.get(0);
+                    opencvService.drawBoundingBoxesOnImage(validBoxes, annotatePath, annotatePath + ".jpg");
                 } catch (Exception e) {
-                    log.error("第{}张图片处理异常: {}", i + 1, path, e);
+                    log.warn("绘制检测框失败", e);
                 }
             }
 
@@ -301,18 +340,25 @@ public class AsyncProcessingService {
             }
 
             // 发送HTTP通知
-            if (targetIp != null && targetPort != null && targetPath != null) {
-                httpNotifyService.sendQrCodeResult(targetIp, targetPort, targetPath,
-                        taskId, result, String.join(",", allImagePaths), allQrCodeResults, scanType);
-            } else {
-                log.warn("多图片异步任务完成，但未配置目标服务器信息");
-            }
+            sendQrCodeResult(targetIp, targetPort, targetPath,
+                    taskId, result, annotatedImageUrl, allQrCodeResults, scanType);
 
         } catch (Exception e) {
             log.error("多图片异步任务执行异常", e);
         }
     }
 
+    private void sendQrCodeResult(String targetIp, Integer targetPort, String targetPath,
+                                   String taskId, String result, String imagePath,
+                                   List<String> qrCodeResults, String scanType) {
+        if (targetIp != null && targetPort != null && targetPath != null) {
+            httpNotifyService.sendQrCodeResult(targetIp, targetPort, targetPath,
+                    taskId, result, imagePath, qrCodeResults, scanType);
+        } else {
+            log.warn("多图片异步任务完成，但未配置目标服务器信息");
+        }
+    }
+
     public static void main(String[] args) {
         String path = "D:\\data\\media\\2026-05-15\\1\\2026-05-15-00-07-20-909.png_3.jpg";
         BufferedImage croppedImage  = null;
@@ -405,67 +451,67 @@ public class AsyncProcessingService {
         if (result != null) return result;
         
         // 2. 尝试灰度化
-        try {
-            log.info("DEBUG: 灰度化开始, image类型={}, 尺寸={}x{}", image.getType(), image.getWidth(), image.getHeight());
-            Mat mat = opencvService.bufferedImageToMat(image);
-            log.info("DEBUG: bufferedImageToMat完成, mat类型={}, channels={}, depth={}", 
-                mat.type(), mat.channels(), mat.depth());
-            Mat gray = opencvService.toGrayscale(mat);
-            log.info("DEBUG: toGrayscale完成, gray channels={}, depth={}", gray.channels(), gray.depth());
-            BufferedImage grayImage = opencvService.matToBufferedImage(gray);
-            log.info("DEBUG: matToBufferedImage完成, grayImage类型={}", grayImage.getType());
-            mat.release();
-            gray.release();
-            result = tryDecode(grayImage, confidence, retryIndex, "灰度");
-            if (result != null) return result;
-        } catch (Exception e) {
-            log.error("灰度预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
-        }
-        
-        // 3. 尝试CLAHE增强
-        try {
-            log.info("DEBUG: CLAHE开始");
-            Mat mat = opencvService.bufferedImageToMat(image);
-            Mat clahe = opencvService.clahe(mat);
-            log.info("DEBUG: clahe完成, clahe channels={}, depth={}", clahe.channels(), clahe.depth());
-            BufferedImage claheImage = opencvService.matToBufferedImage(clahe);
-            mat.release();
-            clahe.release();
-            result = tryDecode(claheImage, confidence, retryIndex, "CLAHE");
-            if (result != null) return result;
-        } catch (Exception e) {
-            log.error("CLAHE预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
-        }
-        
-        // 4. 尝试二值化
-        try {
-            log.info("DEBUG: 二值化开始");
-            Mat mat = opencvService.bufferedImageToMat(image);
-            Mat binary = opencvService.adaptiveThreshold(mat);
-            log.info("DEBUG: adaptiveThreshold完成, binary channels={}, depth={}", binary.channels(), binary.depth());
-            BufferedImage binaryImage = opencvService.matToBufferedImage(binary);
-            mat.release();
-            binary.release();
-            result = tryDecode(binaryImage, confidence, retryIndex, "二值化");
-            if (result != null) return result;
-        } catch (Exception e) {
-            log.error("二值化预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
-        }
-        
-        // 5. 尝试组合预处理（去噪+灰度+CLAHE+二值化）
-        try {
-            log.info("DEBUG: 组合预处理开始");
-            Mat mat = opencvService.bufferedImageToMat(image);
-            Mat processed = opencvService.preprocessForQrCode(mat);
-            log.info("DEBUG: preprocessForQrCode完成, processed channels={}, depth={}", processed.channels(), processed.depth());
-            BufferedImage processedImage = opencvService.matToBufferedImage(processed);
-            mat.release();
-            processed.release();
-            result = tryDecode(processedImage, confidence, retryIndex, "组合预处理");
-            if (result != null) return result;
-        } catch (Exception e) {
-            log.error("组合预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
-        }
+//        try {
+//            log.info("DEBUG: 灰度化开始, image类型={}, 尺寸={}x{}", image.getType(), image.getWidth(), image.getHeight());
+//            Mat mat = opencvService.bufferedImageToMat(image);
+//            log.info("DEBUG: bufferedImageToMat完成, mat类型={}, channels={}, depth={}",
+//                mat.type(), mat.channels(), mat.depth());
+//            Mat gray = opencvService.toGrayscale(mat);
+//            log.info("DEBUG: toGrayscale完成, gray channels={}, depth={}", gray.channels(), gray.depth());
+//            BufferedImage grayImage = opencvService.matToBufferedImage(gray);
+//            log.info("DEBUG: matToBufferedImage完成, grayImage类型={}", grayImage.getType());
+//            mat.release();
+//            gray.release();
+//            result = tryDecode(grayImage, confidence, retryIndex, "灰度");
+//            if (result != null) return result;
+//        } catch (Exception e) {
+//            log.error("灰度预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
+//        }
+//
+//        // 3. 尝试CLAHE增强
+//        try {
+//            log.info("DEBUG: CLAHE开始");
+//            Mat mat = opencvService.bufferedImageToMat(image);
+//            Mat clahe = opencvService.clahe(mat);
+//            log.info("DEBUG: clahe完成, clahe channels={}, depth={}", clahe.channels(), clahe.depth());
+//            BufferedImage claheImage = opencvService.matToBufferedImage(clahe);
+//            mat.release();
+//            clahe.release();
+//            result = tryDecode(claheImage, confidence, retryIndex, "CLAHE");
+//            if (result != null) return result;
+//        } catch (Exception e) {
+//            log.error("CLAHE预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
+//        }
+//
+//        // 4. 尝试二值化
+//        try {
+//            log.info("DEBUG: 二值化开始");
+//            Mat mat = opencvService.bufferedImageToMat(image);
+//            Mat binary = opencvService.adaptiveThreshold(mat);
+//            log.info("DEBUG: adaptiveThreshold完成, binary channels={}, depth={}", binary.channels(), binary.depth());
+//            BufferedImage binaryImage = opencvService.matToBufferedImage(binary);
+//            mat.release();
+//            binary.release();
+//            result = tryDecode(binaryImage, confidence, retryIndex, "二值化");
+//            if (result != null) return result;
+//        } catch (Exception e) {
+//            log.error("二值化预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
+//        }
+//
+//        // 5. 尝试组合预处理（去噪+灰度+CLAHE+二值化）
+//        try {
+//            log.info("DEBUG: 组合预处理开始");
+//            Mat mat = opencvService.bufferedImageToMat(image);
+//            Mat processed = opencvService.preprocessForQrCode(mat);
+//            log.info("DEBUG: preprocessForQrCode完成, processed channels={}, depth={}", processed.channels(), processed.depth());
+//            BufferedImage processedImage = opencvService.matToBufferedImage(processed);
+//            mat.release();
+//            processed.release();
+//            result = tryDecode(processedImage, confidence, retryIndex, "组合预处理");
+//            if (result != null) return result;
+//        } catch (Exception e) {
+//            log.error("组合预处理失败: type={}, msg={}", e.getClass().getName(), e.getMessage(), e);
+//        }
         
         return null;
     }
diff --git a/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/QrCode/WeChatDeCode.java b/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/QrCode/WeChatDeCode.java
index 77994a7..0fa27cd 100644
--- a/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/QrCode/WeChatDeCode.java
+++ b/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/QrCode/WeChatDeCode.java
@@ -21,49 +21,75 @@ import javax.imageio.ImageIO;
 
 import static org.bytedeco.opencv.global.opencv_imgcodecs.imread;
 import static org.opencv.core.CvType.CV_8UC3;
-import org.bytedeco.javacpp.Loader;
 
 /**
- *
+ * 微信二维码解码器（单例模式，模型只加载一次）
  */
 public class WeChatDeCode {
 
+    // ============ 单例模型缓存 ============
+    private static volatile WeChatQRCode instance;
+
+    private static synchronized WeChatQRCode getWeChatQRCode() {
+        if (instance == null) {
+            String modelBasePath = "D:/ccd/wechatqrcode/";
+
+            // 优先使用 _2021nov 版本
+            String detectPrototxt = findModel(modelBasePath, "detect_2021nov.prototxt", "detect.prototxt");
+            String detectCaffemodel = findModel(modelBasePath, "detect_2021nov.caffemodel", "detect.caffemodel");
+            String srPrototxt = findModel(modelBasePath, "sr_2021nov.prototxt", "sr.prototxt");
+            String srCaffemodel = findModel(modelBasePath, "sr_2021nov.caffemodel", "sr.caffemodel");
+
+            validateModel(detectPrototxt, "detect prototxt");
+            validateModel(detectCaffemodel, "detect caffemodel");
+            validateModel(srPrototxt, "sr prototxt");
+            validateModel(srCaffemodel, "sr caffemodel");
+
+            instance = new WeChatQRCode(detectPrototxt, detectCaffemodel, srPrototxt, srCaffemodel);
+            System.out.println("WeChatQRCode 模型加载成功（单例）");
+        }
+        return instance;
+    }
+
+    private static String findModel(String basePath, String preferred, String fallback) {
+        String path = basePath + preferred;
+        if (new File(path).exists()) return path;
+        return basePath + fallback;
+    }
+
+    private static void validateModel(String path, String name) {
+        if (!new File(path).exists()) {
+            throw new RuntimeException("WeChatQRCode 模型文件不存在: " + name + " (" + path + ")");
+        }
+    }
+    // ============ 单例模型缓存 end ============
+
     public static void main(String... args) {
         String imagePath = "D:\\data\\media\\2026-05-20\\1\\2026-05-20-12-20-05-578_1.png_crop_1_6.jpg";
         Mat img = imread(imagePath);
         System.out.println(deCode(img));
-
-        //下载二维码到本地识别
-//        String url="";
-//        System.out.println(openPic(url));
     }
 
     public static String openPic(String url) {
-
         String savePath = "d:\\img";
         String filename = UUID.randomUUID().toString() + ".jpg";
-        String re="0";
+        String re = "0";
         try {
             dxz(url, savePath, filename);
             Mat img = imread(savePath + "\\" + filename);
-            re=deCode(img);
-            //如果识别失败那么尝试放大缩小图片尝试识别，提高准确率
-            if(re.equals("0")) {
+            re = deCode(img);
+            if (re.equals("0")) {
                 ys(savePath, filename);
                 Mat img1 = imread(savePath + "\\1_" + filename);
-                re=deCode(img1);
+                re = deCode(img1);
             }
-            if(re.equals("0")) {
+            if (re.equals("0")) {
                 fd(savePath, filename);
                 Mat img2 = imread(savePath + "\\2_" + filename);
-                re=deCode(img2);
+                re = deCode(img2);
             }
             return re;
-
-
-
         } catch (Exception e) {
-            // TODO Auto-generated catch block
             e.printStackTrace();
             return "0";
         }
@@ -71,13 +97,10 @@ public class WeChatDeCode {
 
     private static void ys(String savePath, String filename) throws Exception {
         try {
-            // 读取原始图片
             BufferedImage image = ImageIO.read(new FileInputStream(savePath + "\\" + filename));
             System.out.println("Width: " + image.getWidth());
             System.out.println("Height: " + image.getHeight());
-            // 调整图片大小
-            BufferedImage newImage = resizeImage(image, image.getWidth()*2, image.getHeight()*2);
-            // 图像缓冲区图片保存为图片文件(文件不存在会自动创建文件保存，文件存在会覆盖原文件保存)
+            BufferedImage newImage = resizeImage(image, image.getWidth() * 2, image.getHeight() * 2);
             ImageIO.write(newImage, "jpg", new File(savePath + "\\1_" + filename));
         } catch (Exception e) {
             e.printStackTrace();
@@ -86,13 +109,10 @@ public class WeChatDeCode {
 
     private static void fd(String savePath, String filename) throws Exception {
         try {
-            // 读取原始图片
             BufferedImage image = ImageIO.read(new FileInputStream(savePath + "\\" + filename));
             System.out.println("Width: " + image.getWidth());
             System.out.println("Height: " + image.getHeight());
-            // 调整图片大小
             BufferedImage newImage = resizeImage(image, 200, 200);
-            // 图像缓冲区图片保存为图片文件(文件不存在会自动创建文件保存，文件存在会覆盖原文件保存)
             ImageIO.write(newImage, "jpg", new File(savePath + "\\2_" + filename));
         } catch (Exception e) {
             e.printStackTrace();
@@ -100,24 +120,17 @@ public class WeChatDeCode {
     }
 
     private static void dxz(String urlString, String savePath, String filename) throws Exception {
-        // 构造URL
         URL url = new URL(urlString);
-        // 打开连接
         URLConnection con = url.openConnection();
-        // 设置请求超时为20s
         con.setConnectTimeout(20 * 1000);
-        // 文件路径不存在 则创建
         File sf = new File(savePath);
         if (!sf.exists()) {
             sf.mkdirs();
         }
-        // jdk 1.7 新特性自动关闭
         try (InputStream in = con.getInputStream();
              OutputStream out = new FileOutputStream(sf.getPath() + "\\" + filename)) {
-            // 创建缓冲区
             byte[] buff = new byte[1024];
             int n;
-            // 开始读取
             while ((n = in.read(buff)) >= 0) {
                 out.write(buff, 0, n);
             }
@@ -125,6 +138,7 @@ public class WeChatDeCode {
             e.printStackTrace();
         }
     }
+
     public static String deCode(BufferedImage croppedImage) {
         Mat img = bufferedImageToMat(croppedImage);
         try {
@@ -140,146 +154,52 @@ public class WeChatDeCode {
         if (image == null) {
             return null;
         }
-
         int width = image.getWidth();
         int height = image.getHeight();
-
-        // 创建 Mat 对象
         Mat mat = new Mat(height, width, CV_8UC3);
-
-        // 获取像素数据
         int[] pixels = new int[width * height];
         image.getRGB(0, 0, width, height, pixels, 0, width);
-
-        // 将像素数据复制到 Mat
         byte[] data = new byte[width * height * 3];
         for (int i = 0; i < width * height; i++) {
             int pixel = pixels[i];
             int r = (pixel >> 16) & 0xFF;
             int g = (pixel >> 8) & 0xFF;
             int b = pixel & 0xFF;
-
-            // OpenCV 使用 BGR 格式
             data[i * 3] = (byte) b;
             data[i * 3 + 1] = (byte) g;
             data[i * 3 + 2] = (byte) r;
         }
-
         mat.data().put(data);
         return mat;
     }
-    private static String getModelPath(String filename) {
-        // 对于 detect 和 sr 文件，优先尝试 _2021nov 版本
-        String altFilename = filename
-            .replace("detect.", "detect_2021nov.")
-            .replace("sr.", "sr_2021nov.");
-        
-        // 模型文件路径（不含中文）
-        String modelBasePath = "D:/ccd/wechatqrcode/";
-        File altFile = new File(modelBasePath + altFilename);
-        File origFile = new File(modelBasePath + filename);
-        
-        if (altFile.exists()) {
-            System.out.println("Loading model: " + altFile.getAbsolutePath());
-            return altFile.getAbsolutePath();
-        } else if (origFile.exists()) {
-            System.out.println("Loading model: " + origFile.getAbsolutePath());
-            return origFile.getAbsolutePath();
-        }
-        
-        System.out.println("Model file not found: " + modelBasePath + altFilename);
-        return modelBasePath + altFilename;
-    }
 
     public static String deCode(Mat img) {
-        // 获取模型文件路径
-        String detectPrototxt = getModelPath("detect_2021nov.prototxt");
-        String detectCaffemodel = getModelPath("detect_2021nov.caffemodel");
-        String srPrototxt = getModelPath("sr_2021nov.prototxt");
-        String srCaffemodel = getModelPath("sr_2021nov.caffemodel");
-        
-        // 验证所有文件都存在
-        if (!new File(detectPrototxt).exists()) {
-            throw new RuntimeException("文件不存在: " + detectPrototxt);
-        }
-        if (!new File(detectCaffemodel).exists()) {
-            throw new RuntimeException("文件不存在: " + detectCaffemodel);
-        }
-        if (!new File(srPrototxt).exists()) {
-            throw new RuntimeException("文件不存在: " + srPrototxt);
-        }
-        if (!new File(srCaffemodel).exists()) {
-            throw new RuntimeException("文件不存在: " + srCaffemodel);
-        }
-        
-        System.out.println("=== 模型文件路径 ===");
-        System.out.println("detectPrototxt: " + detectPrototxt + " (" + new File(detectPrototxt).length() + " bytes)");
-        System.out.println("detectCaffemodel: " + detectCaffemodel + " (" + new File(detectCaffemodel).length() + " bytes)");
-        System.out.println("srPrototxt: " + srPrototxt + " (" + new File(srPrototxt).length() + " bytes)");
-        System.out.println("srCaffemodel: " + srCaffemodel + " (" + new File(srCaffemodel).length() + " bytes)");
-        System.out.println("====================");
-        
-        // 微信二维码对象，需要传入模型文件路径
-        WeChatQRCode we = new WeChatQRCode(detectPrototxt, detectCaffemodel, srPrototxt, srCaffemodel);
-        
-        // 使用 MatVector 存储坐标点
+        WeChatQRCode we = getWeChatQRCode();
         org.bytedeco.opencv.opencv_core.MatVector points = new org.bytedeco.opencv.opencv_core.MatVector();
-        
-        // 微信二维码引擎解码，返回的valList中存放的是解码后的数据，points中存放的是二维码4个角的坐标
         StringVector stringVector = we.detectAndDecode(img, points);
 
         if (stringVector.empty()) {
-            System.out.println("未识别到二维码");
             return "0";
         }
 
-        // 获取第一个二维码内容，并去掉末尾的 null 字符
         String qrContent = stringVector.get(0).getString(StandardCharsets.UTF_8);
-        // 去掉末尾的 null 字符
         if (qrContent != null) {
             qrContent = qrContent.replaceAll("\\x00+$", "");
         }
-
-        // 输出所有识别到的二维码信息
-        System.out.println("========== 识别结果汇总 ==========");
-        System.out.println("共识别到 " + stringVector.size() + " 个二维码");
-        System.out.println("二维码 1: " + qrContent);
-
-        // 输出第一个二维码四个角的坐标
-        if (points.size() > 0) {
-            Mat pointMat = points.get(0);
-            System.out.println("  坐标位置:");
-            for (int j = 0; j < pointMat.rows(); j++) {
-                float x = pointMat.ptr(j).getFloat(0);
-                float y = pointMat.ptr(j).getFloat(4);
-                System.out.println("    角" + (j + 1) + ": (" + x + ", " + y + ")");
-            }
-        }
-        System.out.println("==================================");
-
         return qrContent;
     }
 
     /**
      * 调整图片大小
-     * @param originalImage 原始图片
-     * @param targetWidth 目标宽度
-     * @param targetHeight 目标高度
-     * @return 调整大小后的图片
      */
     private static BufferedImage resizeImage(BufferedImage originalImage, int targetWidth, int targetHeight) {
         BufferedImage resizedImage = new BufferedImage(targetWidth, targetHeight, BufferedImage.TYPE_INT_RGB);
         java.awt.Graphics2D g2d = resizedImage.createGraphics();
-        
-        // 设置渲染质量
         g2d.setRenderingHint(java.awt.RenderingHints.KEY_INTERPOLATION, java.awt.RenderingHints.VALUE_INTERPOLATION_BILINEAR);
         g2d.setRenderingHint(java.awt.RenderingHints.KEY_RENDERING, java.awt.RenderingHints.VALUE_RENDER_QUALITY);
         g2d.setRenderingHint(java.awt.RenderingHints.KEY_ANTIALIASING, java.awt.RenderingHints.VALUE_ANTIALIAS_ON);
-        
-        // 绘制缩放后的图片
         g2d.drawImage(originalImage, 0, 0, targetWidth, targetHeight, null);
         g2d.dispose();
-        
         return resizedImage;
     }
 
diff --git a/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/camera/hik/ImageCaptureService.java b/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/camera/hik/ImageCaptureService.java
index 312446c..af4c087 100644
--- a/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/camera/hik/ImageCaptureService.java
+++ b/src/main/java/com/example/lxcameraapi/service/IndustrialCamera/camera/hik/ImageCaptureService.java
@@ -31,6 +31,12 @@ public class ImageCaptureService {
     // SDK初始化状态
     private boolean isSdkInitialized = false;
 
+    // 相机重连最大重试次数
+    private static final int RECONNECT_MAX_RETRIES = 3;
+
+    // 相机重连间隔(ms)
+    private static final int RECONNECT_INTERVAL = 3000;
+
     @Resource
     private AppConfig appConfig;
     /**
@@ -242,6 +248,17 @@ public class ImageCaptureService {
      * @return 是否成功
      */
     public boolean captureImage(String ip, String savePath) {
+        return captureImage(ip, savePath, false);
+    }
+
+    /**
+     * 根据IP拍照并保存
+     * @param ip 相机IP
+     * @param savePath 保存路径
+     * @param isRetry 是否为重试（避免无限递归）
+     * @return 是否成功
+     */
+    private boolean captureImage(String ip, String savePath, boolean isRetry) {
         if (!isSdkInitialized) {
             log.error("SDK未初始化");
             return false;
@@ -251,6 +268,13 @@ public class ImageCaptureService {
         Handle hCamera = handleMap.get(ip);
         if (hCamera == null) {
             log.error("相机 {} 未初始化，请先调用init方法", ip);
+            // 未初始化时也尝试重连
+            if (!isRetry) {
+                Handle newHandle = reconnectCamera(ip);
+                if (newHandle != null) {
+                    return captureImage(ip, savePath, true);
+                }
+            }
             return false;
         }
 
@@ -262,6 +286,12 @@ public class ImageCaptureService {
             int nRet = MvCameraControl.MV_CC_GetIntValue(hCamera, "PayloadSize", stParam);
             if (MV_OK != nRet) {
                 log.error("获取PayloadSize失败，errcode: [0x{}]", Integer.toHexString(nRet));
+                if (!isRetry) {
+                    Handle newHandle = reconnectCamera(ip);
+                    if (newHandle != null) {
+                        return captureImage(ip, savePath, true);
+                    }
+                }
                 return false;
             }
 
@@ -272,6 +302,12 @@ public class ImageCaptureService {
 
             if (MV_OK != nRet) {
                 log.error("获取图像失败，errcode: [0x{}]", Integer.toHexString(nRet));
+                if (!isRetry) {
+                    Handle newHandle = reconnectCamera(ip);
+                    if (newHandle != null) {
+                        return captureImage(ip, savePath, true);
+                    }
+                }
                 return false;
             }
 
@@ -283,6 +319,12 @@ public class ImageCaptureService {
 
         } catch (Exception e) {
             log.error("拍照异常", e);
+            if (!isRetry) {
+                Handle newHandle = reconnectCamera(ip);
+                if (newHandle != null) {
+                    return captureImage(ip, savePath, true);
+                }
+            }
             return false;
         }
     }
@@ -405,6 +447,69 @@ public class ImageCaptureService {
         return handleMap.containsKey(ip);
     }
 
+    /**
+     * 销毁单个相机的句柄并从map中移除
+     */
+    private void destroyCameraHandle(String ip) {
+        Handle hCamera = handleMap.remove(ip);
+        if (hCamera == null) return;
+        try {
+            MvCameraControl.MV_CC_StopGrabbing(hCamera);
+        } catch (Exception e) {
+            log.warn("相机 {} 停止抓图异常", ip, e);
+        }
+        try {
+            MvCameraControl.MV_CC_CloseDevice(hCamera);
+        } catch (Exception e) {
+            log.warn("相机 {} 关闭设备异常", ip, e);
+        }
+        try {
+            MvCameraControl.MV_CC_DestroyHandle(hCamera);
+        } catch (Exception e) {
+            log.warn("相机 {} 销毁句柄异常", ip, e);
+        }
+        log.info("相机 {} 句柄已销毁", ip);
+    }
+
+    /**
+     * 重连相机，带重试
+     * @return 新句柄，失败返回null
+     */
+    private Handle reconnectCamera(String ip) {
+        log.warn("相机 {} 开始重连...", ip);
+        destroyCameraHandle(ip);
+
+        for (int i = 1; i <= RECONNECT_MAX_RETRIES; i++) {
+            try {
+                // 重新枚举设备
+                ArrayList<MV_CC_DEVICE_INFO> stDeviceList = MV_CC_EnumDevices(MV_GIGE_DEVICE | MV_USB_DEVICE);
+                if (stDeviceList == null || stDeviceList.isEmpty()) {
+                    log.warn("相机 {} 重连失败（第{}/{}次），未枚举到设备", ip, i, RECONNECT_MAX_RETRIES);
+                } else {
+                    Handle newHandle = initSingleCamera(ip, stDeviceList);
+                    if (newHandle != null) {
+                        log.info("相机 {} 重连成功（第{}次尝试）", ip, i);
+                        return newHandle;
+                    }
+                    log.warn("相机 {} 重连失败（第{}/{}次），初始化句柄失败", ip, i, RECONNECT_MAX_RETRIES);
+                }
+            } catch (Exception e) {
+                log.warn("相机 {} 重连异常（第{}/{}次）", ip, i, RECONNECT_MAX_RETRIES, e);
+            }
+
+            if (i < RECONNECT_MAX_RETRIES) {
+                try {
+                    Thread.sleep(RECONNECT_INTERVAL);
+                } catch (InterruptedException e) {
+                    Thread.currentThread().interrupt();
+                    break;
+                }
+            }
+        }
+        log.error("相机 {} 重连失败，已达最大重试次数 {}", ip, RECONNECT_MAX_RETRIES);
+        return null;
+    }
+
     /**
      * 连续拍照
      * @param ip 相机IP
diff --git a/src/main/resources/application.yml b/src/main/resources/application.yml
index 3067d1d..976fac1 100644
--- a/src/main/resources/application.yml
+++ b/src/main/resources/application.yml
@@ -52,7 +52,7 @@ deleteFile:
    - "D://go/deleteFile"
   day: 7
 picPath: "D:/data/media/"
-picUrl: "http://127.0.0.1:9012/pic/"
+picUrl: "http://192.168.100.110:9012/pic/"
 
 serve:
   ip: 127.0.0.1
@@ -76,6 +76,6 @@ yoloModelConfig:
   #  图片大小
       imgSize: 2048
   #  置信度
-      confThreshold: 0.90
+      confThreshold: 0.9
       names: ['code']