如何在Android Studio中解析致命信号11（SIGSEGV），代码1（SEGV_MAPERR）

Question

我尝试在 Android Studio 应用程序中使用 .tflite 文件。该模型能够检测图像中的形状。目标是围绕感兴趣的形状创建边界框。为此我制作了这个程序：

    private MappedByteBuffer loadModelFile() throws IOException {
        AssetFileDescriptor fileDescriptor = getContext().getAssets().openFd("yolov4.tflite");
        FileInputStream inputStream = new FileInputStream(fileDescriptor.getFileDescriptor());
        FileChannel fileChannel = inputStream.getChannel();
        long startOffset = fileDescriptor.getStartOffset();
        long declaredLength = fileDescriptor.getDeclaredLength();
        MappedByteBuffer modelBuffer = fileChannel.map(FileChannel.MapMode.READ_ONLY, startOffset, 
declaredLength);

        if (modelBuffer != null) {
            System.out.println("Model load with success.");
        } else {
            System.out.println("Error during loading model.");
        }

        return modelBuffer;
    }

    // Method to do the inference
    public void createBoundingBox(Bitmap image){

        int width = image.getWidth();
        int height = image.getHeight();
        Bitmap resizedImage = Bitmap.createBitmap(width, height, image.getConfig());
        try {
            MappedByteBuffer tfliteModel = loadModelFile();
            // Creates inputs for reference.
            TensorBuffer inputFeature0 = TensorBuffer.createFixedSize(new int[]{1, 416, 416, 3}, DataType.FLOAT32);

            ByteBuffer byteBuffer = convertBitmapToByteBuffer(resizedImage);

            inputFeature0.loadBuffer(byteBuffer);

            Interpreter tflite = new Interpreter(tfliteModel);

            Object[] inputs = {inputFeature0.getBuffer()};
            Map<Integer, Object> outputs = new HashMap<>();
            TensorBuffer outputFeature0 = TensorBuffer.createFixedSize(new int[]{1, 416, 416, 3}, DataType.FLOAT32);
            outputs.put(0, outputFeature0.getBuffer());

            tflite.runForMultipleInputsOutputs(inputs, outputs);

            List<Detection> detections = processModelOutputs(outputFeature0);

            //draw the bounding box
            Canvas canvas = new Canvas(resizedImage);
            Paint paint = new Paint();
            paint.setStyle(Paint.Style.STROKE);
            paint.setColor(Color.RED);
            paint.setStrokeWidth(2);

            for(Detection detection : detections){
                float left = (detection.centerX - detection.width/2)*resizedImage.getWidth();
                float top = (detection.centerY - detection.height/2)*resizedImage.getHeight();
                float bottom = (detection.centerY + detection.height/2)*resizedImage.getHeight();
                float right = (detection.centerX + detection.width/2)*resizedImage.getWidth();
                canvas.drawRect(left, top, right, bottom, paint);
            }

            imageView.setImageBitmap(resizedImage);

            // Releases model resources if no longer used.
            tflite.close();
        } catch (IOException e) {
            System.out.println("Not connected to the model");
        }
    }

private ByteBuffer convertBitmapToByteBuffer(Bitmap bitmap){
        ByteBuffer byteBuffer = ByteBuffer.allocateDirect(4*416*416*3);
        byteBuffer.order(ByteOrder.nativeOrder());
        int[] intValues = new int[imageSize*imageSize];
        bitmap.getPixels(intValues, 0, 416, 0, 0, 416, 416);
        for (int i = 0; i<imageSize; i++){
            for(int j = 0; j<imageSize; j++) {
                int val = intValues[i*416+j];
                byteBuffer.putFloat(((val>>16)&0xFF)*(1.f/255));
                byteBuffer.putFloat(((val>>8)&0xFF)*(1.f/255));
                byteBuffer.putFloat((val&0xFF)*(1.f/255));
            }
        }

        return byteBuffer;
    }

然而，此行出现错误：“致命信号 11 (SIGSEGV)，代码 1 (SEGV_MAPERR)”：“outputs.put(0, outputFeature0.getBuffer());”。

我尝试使用yolov4.tflite模型的java代码

try {
    Yolov4 model = Yolov4.newInstance(context);

    // Creates inputs for reference.
    TensorBuffer inputFeature0 = TensorBuffer.createFixedSize(new int[]{1, 416, 416, 3}, DataType.FLOAT32);
    inputFeature0.loadBuffer(byteBuffer);

    // Runs model inference and gets result.
    Yolov4.Outputs outputs = model.process(inputFeature0);
    TensorBuffer outputFeature0 = outputs.getOutputFeature0AsTensorBuffer();

    // Releases model resources if no longer used.
    model.close();
} catch (IOException e) {
    // TODO Handle the exception
}

但是“TensorBuffer inputFeature0 = TensorBuffer.createFixedSize(new int[]{1, 416, 416, 3}, DataType.FLOAT32);”行出现了同样的错误

Answer 1

马031-83-7892 马 034076669 马 030-34-4941 电话 137351341 洛杉矶 509847852 弗吉尼亚州 229509583 弗吉尼亚州 226-64-7327 弗吉尼亚州 227327199 弗吉尼亚州 229-61-0501 弗吉尼亚州 23011070966 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 VA 2301-50 -12525 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据 无辜数据