Android集成ARCore与ML Kit调用process方法报内部错误及SIGSEGV崩溃

我正按照ARCore官方文档的指引，在Android Studio中开展ARCore与ML Kit的集成开发，所使用的Java代码如下：

ObjectDetector objectDetector = createObjectDetector();
  Image cameraImage = null;
  try {
    cameraImage = frame.acquireCameraImage();
        // Process `cameraImage` using your ML inference model.
        objectDetector.process(InputImage.fromMediaImage(cameraImage, getRotationCompensation("1", this, false)))
                .addOnSuccessListener(
                        new OnSuccessListener<List<DetectedObject>>() {
                          @Override
                          public void onSuccess(List<DetectedObject> detectedObjects) {
                            // Task completed successfully
                            // The list of detected objects contains one item if multiple
                            // object detection wasn't enabled.
                            for (DetectedObject detectedObject : detectedObjects) {
                            Rect boundingBox = detectedObject.getBoundingBox();
                            Integer trackingId = detectedObject.getTrackingId();
                            for (DetectedObject.Label label : detectedObject.getLabels()) {
                              String text = label.getText();
                              if (PredefinedCategory.FOOD.equals(text)) {
                              }
                              int index = label.getIndex();
                              if (PredefinedCategory.FOOD_INDEX == index) {
                              }
                              float confidence = label.getConfidence();
                            }
                            }
                          }
                        })
                .addOnFailureListener(
                        new OnFailureListener() {
                          @Override
                          public void onFailure(@NonNull Exception e) {
                            // Task failed with an exception
                            Log.e("Failure Listener", e.getMessage());
                          }
                        });
      } catch (NotYetAvailableException e) {
        // NotYetAvailableException is an exception that can be expected when the camera is not ready
        // yet. The image may become available on a next frame.
      } catch (RuntimeException e) {
        // A different exception occurred, e.g. DeadlineExceededException, ResourceExhaustedException.
        // Handle this error appropriately.
        Log.e("Error", e.toString());
      } catch (CameraAccessException e) {
        e.printStackTrace();
      } finally {
        if (cameraImage != null) {
          cameraImage.close();
        }
      }

代码运行时出现如下报错：

Internal error has occurred when executing ML Kit tasks
Fatal signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x4faee70000000 in tid 14906 ...

经初步定位，错误由objectDetector.process(...)触发。

这个SIGSEGV段错误属于native层内存访问违规，核心原因是图像资源提前释放：
ML Kit的process()方法是异步执行的，调用该方法后推理任务会在后台线程运行，当前线程会立刻走到外层finally代码块执行cameraImage.close()。此时后台推理线程还在读取cameraImage绑定的native内存，访问已经被释放的内存地址就会触发段错误崩溃。

另外两个可能触发同类崩溃的诱因：

• ObjectDetector实例频繁创建销毁，没有全局复用，导致native内存泄漏/野指针
• getRotationCompensation()返回的旋转角度不是0/90/180/270四个合法值，触发ML Kit图像处理逻辑的native异常

• 移除原有外层try-catch的finally块中关闭cameraImage的逻辑
• 将图像关闭逻辑绑定到ML Kit推理任务的完成回调中，保证推理完成后再释放资源
• 将ObjectDetector改为全局单例复用，不要每帧新建检测器
• 校验旋转角度返回值，确保为0、90、180、270四个合法值

修正后的核心代码如下：

// ObjectDetector设为全局单例复用，禁止在帧回调中重复创建
private final ObjectDetector objectDetector = createObjectDetector();

// 帧处理逻辑
Image cameraImage = null;
try {
    cameraImage = frame.acquireCameraImage();
    int rotation = getRotationCompensation("1", this, false);
    // 校验旋转角度合法性
    if (rotation % 90 != 0) {
        cameraImage.close();
        return;
    }
    InputImage inputImage = InputImage.fromMediaImage(cameraImage, rotation);
    objectDetector.process(inputImage)
            .addOnSuccessListener(detectedObjects -> {
                // 原有检测结果处理逻辑保持不变
                for (DetectedObject detectedObject : detectedObjects) {
                    Rect boundingBox = detectedObject.getBoundingBox();
                    Integer trackingId = detectedObject.getTrackingId();
                    for (DetectedObject.Label label : detectedObject.getLabels()) {
                        String text = label.getText();
                        if (PredefinedCategory.FOOD.equals(text)) {}
                        int index = label.getIndex();
                        if (PredefinedCategory.FOOD_INDEX == index) {}
                        float confidence = label.getConfidence();
                    }
                }
            })
            .addOnFailureListener(e -> Log.e("ML Kit Inference", e.getMessage()))
            // 推理任务无论成功失败，执行完成后再释放图像资源
            .addOnCompleteListener(task -> {
                if (cameraImage != null) {
                    cameraImage.close();
                }
            });
} catch (NotYetAvailableException e) {
    // 相机未就绪属于正常情况，直接跳过当前帧即可
} catch (RuntimeException | CameraAccessException e) {
    Log.e("Frame Acquire Error", e.toString());
    // 获取帧异常分支也要及时释放已获取的图像
    if (cameraImage != null) {
        cameraImage.close();
    }
}

内容的提问来源于stack exchange，提问作者salam