关于深度学习:Android中Bitmap保持宽高比缩放

需要形容

深度学习模型在 Android 挪动端部署的时候，对于采集到的的摄像头画面须要放弃宽高比的将 Bitmap 缩放到模型输出的大小，缩放后较指标尺寸像素缺失的局部采纳灰度填充形式, 以避免指标产生显著形变影响辨认成果。在本例中，深度模型是以 MobileNetV2 为 Backbone 网络的 YOLOv3，并基于 MNN 挪动端推理框架部署。其中模型输出图像的尺寸大小为 (320,320)。

实现

public static Bitmap scaleImage(Bitmap bm, int newWidth, int newHeight)
{if (bm == null) {return null;}
    int width = bm.getWidth();
    int height = bm.getHeight();
    float scaleWidth = ((float) newWidth) / width;
    float scaleHeight = ((float) newHeight) / height;
    
    // 放弃宽高比缩放，以长边为主
    float scaleRatio = Math.min(scaleHeight, scaleWidth);
    Matrix matrix = new Matrix();
    matrix.postScale(scaleRatio, scaleRatio);
    Bitmap newBm = Bitmap.createBitmap(bm, 0, 0, width, height, matrix, true);
    
    // 创立指标大小 Bitmap
    Bitmap scaledImage = Bitmap.createBitmap(newWidth, newHeight, Bitmap.Config.ARGB_8888);
    Canvas canvas = new Canvas(scaledImage);
    
    // 绘制背景色彩
    Paint paint = new Paint();
    paint.setColor(Color.GRAY);
    paint.setStyle(Paint.Style.FILL);
    canvas.drawRect(0, 0, canvas.getWidth(),    canvas.getHeight(), paint);
    
    // 确定画面地位
    float left = 0;
    float top = 0;
    if (width > height){top = (float)((newBm.getWidth() - newBm.getHeight()) / 2.0);
    }
    else{left = (float)((newBm.getHeight() - newBm.getWidth()) / 2.0);
    }
    canvas.drawBitmap(newBm, left , top, null);
    if (!bm.isRecycled()){bm.recycle();
    }
    return scaledImage;
}

成果

摄像头捕捉到的原始画面 Bitmap:

放弃宽高比缩放后的画面 Bitmap: