Opencv3.1 MultiBandBlender源码分析

void MultiBandBlender::prepare(Rect dst_roi)
{
dst_roi_final_ = dst_roi;

// Crop unnecessary bands
double max_len = static_cast<double>(std::max(dst_roi.width, dst_roi.height));
num_bands_ = std::min(actual_num_bands_, static_cast<int>(ceil(std::log(max_len) / std::log(2.0))));

// Add border to the final image, to ensure sizes are divided by (1 << num_bands_)
dst_roi.width += ((1 << num_bands_) - dst_roi.width % (1 << num_bands_)) % (1 << num_bands_);
dst_roi.height += ((1 << num_bands_) - dst_roi.height % (1 << num_bands_)) % (1 << num_bands_);

Blender::prepare(dst_roi);

dst_pyr_laplace_.resize(num_bands_ + 1);
dst_pyr_laplace_[0] = dst_;

dst_band_weights_.resize(num_bands_ + 1);
dst_band_weights_[0].create(dst_roi.size(), weight_type_);
dst_band_weights_[0].setTo(0);

for (int i = 1; i <= num_bands_; ++i)
{
dst_pyr_laplace_[i].create((dst_pyr_laplace_[i - 1].rows + 1) / 2,
(dst_pyr_laplace_[i - 1].cols + 1) / 2, CV_16SC3);
dst_band_weights_[i].create((dst_band_weights_[i - 1].rows + 1) / 2,
(dst_band_weights_[i - 1].cols + 1) / 2, weight_type_);
dst_pyr_laplace_[i].setTo(Scalar::all(0));
dst_band_weights_[i].setTo(0);
}
}

1. 确定最终ROI区域的大小

2. 确定最终的要用多少个band

3. 对感兴趣区域的宽和高进行调整，确保能被 1 << num_bands_ 整除

4. 调用Blender类下的prepare方法，创建dst_，dst_mask和把 dst_roi赋值给Blender下的成员对象dst_roi_

5. 将ROI区域赋值给dst_pyr_laplace_的第0层，拉普拉斯金字塔总共num_bands_+1 层，相应的band权重的金字塔也为 num_bands_+1 层，band权重金字塔第0层的清零

6. 金字塔第1层的宽和高是第0层的一半，同理类推，第2层是1层的一半，第3层是2层的一半… ，最终权重金字塔的每一层都清零

分析

void MultiBandBlender::feed(InputArray _img, InputArray mask, Point tl)

void MultiBandBlender::feed(InputArray _img, InputArray mask, Point tl)
{
#ifdef ENABLE_LOG
int64 t = getTickCount();
#endif

UMat img = _img.getUMat();
CV_Assert(img.type() == CV_16SC3 || img.type() == CV_8UC3);
CV_Assert(mask.type() == CV_8U);

// Keep source image in memory with small border
int gap = 3 * (1 << num_bands_);
Point tl_new(std::max(dst_roi_.x, tl.x - gap),
std::max(dst_roi_.y, tl.y - gap));
Point br_new(std::min(dst_roi_.br().x, tl.x + img.cols + gap),
std::min(dst_roi_.br().y, tl.y + img.rows + gap));

// Ensure coordinates of top-left, bottom-right corners are divided by (1 << num_bands_).
// After that scale between layers is exactly 2.
//
// We do it to avoid interpolation problems when keeping sub-images only. There is no such problem when
// image is bordered to have size equal to the final image size, but this is too memory hungry approach.
tl_new.x = dst_roi_.x + (((tl_new.x - dst_roi_.x) >> num_bands_) << num_bands_);
tl_new.y = dst_roi_.y + (((tl_new.y - dst_roi_.y) >> num_bands_) << num_bands_);
int width = br_new.x - tl_new.x;
int height = br_new.y - tl_new.y;
width += ((1 << num_bands_) - width % (1 << num_bands_)) % (1 << num_bands_);
height += ((1 << num_bands_) - height % (1 << num_bands_)) % (1 << num_bands_);
br_new.x = tl_new.x + width;
br_new.y = tl_new.y + height;
int dy = std::max(br_new.y - dst_roi_.br().y, 0);
int dx = std::max(br_new.x - dst_roi_.br().x, 0);
tl_new.x -= dx; br_new.x -= dx;
tl_new.y -= dy; br_new.y -= dy;

int top = tl.y - tl_new.y;
int left = tl.x - tl_new.x;
int bottom = br_new.y - tl.y - img.rows;
int right = br_new.x - tl.x - img.cols;

// Create the source image Laplacian pyramid
UMat img_with_border;
copyMakeBorder(_img, img_with_border, top, bottom, left, right,
BORDER_REFLECT);
LOGLN("  Add border to the source image, time: " << ((getTickCount() - t) / getTickFrequency()) << " sec");
#ifdef ENABLE_LOG
t = getTickCount();
#endif

std::vector<UMat> src_pyr_laplace;
if (can_use_gpu_ && img_with_border.depth() == CV_16S)
createLaplacePyrGpu(img_with_border, num_bands_, src_pyr_laplace);
else
createLaplacePyr(img_with_border, num_bands_, src_pyr_laplace);

LOGLN("  Create the source image Laplacian pyramid, time: " << ((getTickCount() - t) / getTickFrequency()) << " sec");
#ifdef ENABLE_LOG
t = getTickCount();
#endif

// Create the weight map Gaussian pyramid
UMat weight_map;
std::vector<UMat> weight_pyr_gauss(num_bands_ + 1);

if(weight_type_ == CV_32F)
{
mask.getUMat().convertTo(weight_map, CV_32F, 1./255.);
}
else // weight_type_ == CV_16S
{
mask.getUMat().convertTo(weight_map, CV_16S);
UMat add_mask;
compare(mask, 0, add_mask, CMP_NE);
add(weight_map, Scalar::all(1), weight_map, add_mask);
}

copyMakeBorder(weight_map, weight_pyr_gauss[0], top, bottom, left, right, BORDER_CONSTANT);

for (int i = 0; i < num_bands_; ++i)
pyrDown(weight_pyr_gauss[i], weight_pyr_gauss[i + 1]);

LOGLN("  Create the weight map Gaussian pyramid, time: " << ((getTickCount() - t) / getTickFrequency()) << " sec");
#ifdef ENABLE_LOG
t = getTickCount();
#endif

int y_tl = tl_new.y - dst_roi_.y;
int y_br = br_new.y - dst_roi_.y;
int x_tl = tl_new.x - dst_roi_.x;
int x_br = br_new.x - dst_roi_.x;

// Add weighted layer of the source image to the final Laplacian pyramid layer
for (int i = 0; i <= num_bands_; ++i)
{
Rect rc(x_tl, y_tl, x_br - x_tl, y_br - y_tl);
#ifdef HAVE_OPENCL
if ( !cv::ocl::useOpenCL() ||
!ocl_MultiBandBlender_feed(src_pyr_laplace[i], weight_pyr_gauss[i],
dst_pyr_laplace_[i](rc), dst_band_weights_[i](rc)) )
#endif
{
Mat _src_pyr_laplace = src_pyr_laplace[i].getMat(ACCESS_READ);
Mat _dst_pyr_laplace = dst_pyr_laplace_[i](rc).getMat(ACCESS_RW);
Mat _weight_pyr_gauss = weight_pyr_gauss[i].getMat(ACCESS_READ);
Mat _dst_band_weights = dst_band_weights_[i](rc).getMat(ACCESS_RW);
if(weight_type_ == CV_32F)
{
for (int y = 0; y < rc.height; ++y)
{
const Point3_<short>* src_row = _src_pyr_laplace.ptr<Point3_<short> >(y);
Point3_<short>* dst_row = _dst_pyr_laplace.ptr<Point3_<short> >(y);
const float* weight_row = _weight_pyr_gauss.ptr<float>(y);
float* dst_weight_row = _dst_band_weights.ptr<float>(y);

for (int x = 0; x < rc.width; ++x)
{
dst_row[x].x += static_cast<short>(src_row[x].x * weight_row[x]);
dst_row[x].y += static_cast<short>(src_row[x].y * weight_row[x]);
dst_row[x].z += static_cast<short>(src_row[x].z * weight_row[x]);
dst_weight_row[x] += weight_row[x];
}
}
}
else // weight_type_ == CV_16S
{
for (int y = 0; y < y_br - y_tl; ++y)
{
const Point3_<short>* src_row = _src_pyr_laplace.ptr<Point3_<short> >(y);
Point3_<short>* dst_row = _dst_pyr_laplace.ptr<Point3_<short> >(y);
const short* weight_row = _weight_pyr_gauss.ptr<short>(y);
short* dst_weight_row = _dst_band_weights.ptr<short>(y);

for (int x = 0; x < x_br - x_tl; ++x)
{
dst_row[x].x += short((src_row[x].x * weight_row[x]) >> 8);
dst_row[x].y += short((src_row[x].y * weight_row[x]) >> 8);
dst_row[x].z += short((src_row[x].z * weight_row[x]) >> 8);
dst_weight_row[x] += weight_row[x];
}
}
}
}
#ifdef HAVE_OPENCL
else
{
CV_IMPL_ADD(CV_IMPL_OCL);
}
#endif

x_tl /= 2; y_tl /= 2;
x_br /= 2; y_br /= 2;
}

LOGLN("  Add weighted layer of the source image to the final Laplacian pyramid layer, time: " << ((getTickCount() - t) / getTickFrequency()) << " sec");
}

1. 分析创建拉普拉斯金字塔源码分析

void createLaplacePyr(InputArray img, int num_levels, std::vector &pyr)

void createLaplacePyr(InputArray img, int num_levels, std::vector<UMat> &pyr)
{
#ifdef HAVE_TEGRA_OPTIMIZATION
cv::Mat imgMat = img.getMat();
if(tegra::useTegra() && tegra::createLaplacePyr(imgMat, num_levels, pyr))
return;
#endif
//金字塔的大小和容量都是num_levels+1
pyr.resize(num_levels + 1);

if(img.depth() == CV_8U)
{
if(num_levels == 0)
{
img.getUMat().convertTo(pyr[0], CV_16S);
return;
}

UMat downNext;
UMat current = img.getUMat();
pyrDown(img, downNext);

for(int i = 1; i < num_levels; ++i)
{
UMat lvl_up;
UMat lvl_down;

pyrDown(downNext, lvl_down);
pyrUp(downNext, lvl_up, current.size());
subtract(current, lvl_up, pyr[i-1], noArray(), CV_16S);

current = downNext;
downNext = lvl_down;
}

{
UMat lvl_up;
pyrUp(downNext, lvl_up, current.size());
subtract(current, lvl_up, pyr[num_levels-1], noArray(), CV_16S);

downNext.convertTo(pyr[num_levels], CV_16S);
}
}
else
{
pyr[0] = img.getUMat();
for (int i = 0; i < num_levels; ++i)
pyrDown(pyr[i], pyr[i + 1]);
UMat tmp;
for (int i = 0; i < num_levels; ++i)
{
pyrUp(pyr[i + 1], tmp, pyr[i].size());
subtract(pyr[i], tmp, pyr[i]);
}
}
}