added Harris corner detector into gpu module

diff --git a/modules/gpu/include/opencv2/gpu/gpu.hpp b/modules/gpu/include/opencv2/gpu/gpu.hpp
index cdc9e4a..043f163 100644 (file)
--- a/modules/gpu/include/opencv2/gpu/gpu.hpp
+++ b/modules/gpu/include/opencv2/gpu/gpu.hpp
@@ -627,6 +627,10 @@ namespace cv
         //! disabled until fix crash\r
         CV_EXPORTS void Canny(const GpuMat& image, GpuMat& edges, double threshold1, double threshold2, int apertureSize = 3);\r
 \r
+        //! computes Harris cornerness criteria at each image pixel \r
+        // (does BORDER_CONSTANT interpolation with 0 as the fill value)\r
+        CV_EXPORTS void cornerHarris(const GpuMat& src, GpuMat& dst, int blockSize, int apertureSize, double k);\r
+\r
         //////////////////////////////// Filter Engine ////////////////////////////////\r
 \r
         /*!\r

diff --git a/modules/gpu/src/cuda/imgproc.cu b/modules/gpu/src/cuda/imgproc.cu
index 2326119..1e632a2 100644 (file)
--- a/modules/gpu/src/cuda/imgproc.cu
+++ b/modules/gpu/src/cuda/imgproc.cu
@@ -463,4 +463,52 @@ namespace cv { namespace gpu { namespace imgproc
     {\r
         reprojectImageTo3D_caller(disp, xyzw, q, stream);\r
     }\r
+\r
+/////////////////////////////////////////// Corner Harris /////////////////////////////////////////////////\r
+\r
+    __global__ void cornerHarris_kernel(const int cols, const int rows, const int block_size, const float k, const PtrStep Dx, const PtrStep Dy, PtrStep dst)\r
+    {\r
+        const unsigned int x = blockIdx.x * blockDim.x + threadIdx.x;\r
+        const unsigned int y = blockIdx.y * blockDim.y + threadIdx.y;\r
+\r
+        if (x < cols && y < rows)\r
+        {\r
+            float a = 0.f;\r
+            float b = 0.f;\r
+            float c = 0.f;\r
+\r
+            const unsigned int j_begin = max(x - block_size, 0);\r
+            const unsigned int i_begin = max(y - block_size, 0);\r
+            const unsigned int j_end = min(x + block_size + 1, cols);\r
+            const unsigned int i_end = min(y + block_size + 1, rows);\r
+\r
+            for (unsigned int i = i_begin; i < i_end; ++i)\r
+            {\r
+                const float* dx_row = (const float*)Dx.ptr(i);\r
+                const float* dy_row = (const float*)Dy.ptr(i);\r
+                for (unsigned int j = j_begin; j < j_end; ++j)\r
+                {\r
+                    float dx = dx_row[j];\r
+                    float dy = dy_row[j];\r
+                    a += dx * dx;\r
+                    b += dx * dy;\r
+                    c += dy * dy;\r
+                }\r
+            }\r
+\r
+            ((float*)dst.ptr(y))[x] = a * c - b * b - k * (a + c) * (a + c);\r
+        }\r
+    }\r
+\r
+    void cornerHarris_caller(const int block_size, const float k, const DevMem2D Dx, const DevMem2D Dy, DevMem2D dst)\r
+    {\r
+        const int rows = Dx.rows;\r
+        const int cols = Dx.cols;\r
+\r
+        dim3 threads(32, 8);\r
+        dim3 grid(divUp(cols, threads.x), divUp(rows, threads.y));\r
+\r
+        cornerHarris_kernel<<<grid, threads>>>(cols, rows, block_size / 2, k, Dx, Dy, dst);\r
+        cudaSafeCall(cudaThreadSynchronize());\r
+    }\r
 }}}\r

diff --git a/modules/gpu/src/imgproc_gpu.cpp b/modules/gpu/src/imgproc_gpu.cpp
index 74f3014..00696d8 100644 (file)
--- a/modules/gpu/src/imgproc_gpu.cpp
+++ b/modules/gpu/src/imgproc_gpu.cpp
@@ -68,6 +68,8 @@ void cv::gpu::histEven(const GpuMat&, GpuMat&, int, int, int) { throw_nogpu(); }
 void cv::gpu::histEven(const GpuMat&, GpuMat*, int*, int*, int*) { throw_nogpu(); }\r
 void cv::gpu::histRange(const GpuMat&, GpuMat&, const GpuMat&) { throw_nogpu(); }\r
 void cv::gpu::histRange(const GpuMat&, GpuMat*, const GpuMat*) { throw_nogpu(); }\r
+void cv::gpu::cornerHarris(const GpuMat&, GpuMat&, int, int, double) { throw_nogpu(); }\r
+\r
 \r
 #else /* !defined (HAVE_CUDA) */\r
 \r
@@ -856,4 +858,35 @@ void cv::gpu::histRange(const GpuMat& src, GpuMat hist[4], const GpuMat levels[4
     hist_callers[src.depth()](src, hist, levels);\r
 }\r
 \r
+namespace cv { namespace gpu { namespace imgproc {\r
+\r
+    void cornerHarris_caller(const int block_size, const float k, const DevMem2D Dx, const DevMem2D Dy, DevMem2D dst);\r
+\r
+}}}\r
+\r
+void cv::gpu::cornerHarris(const GpuMat& src, GpuMat& dst, int blockSize, int apertureSize, double k)\r
+{\r
+    CV_Assert(src.type() == CV_32F);\r
+\r
+    double scale = (double)(1 << ((apertureSize > 0 ? apertureSize : 3) - 1)) * blockSize;\r
+    if (apertureSize < 0) scale *= 2.;\r
+    scale = 1./scale;\r
+\r
+    GpuMat Dx, Dy;\r
+    if (apertureSize > 0)\r
+    {\r
+        Sobel(src, Dx, CV_32F, 1, 0, apertureSize, scale);\r
+        Sobel(src, Dy, CV_32F, 0, 1, apertureSize, scale);\r
+    }\r
+    else\r
+    {\r
+        Scharr(src, Dx, CV_32F, 1, 0, scale);\r
+        Scharr(src, Dy, CV_32F, 0, 1, scale);\r
+    }\r
+\r
+    dst.create(src.size(), CV_32F);\r
+    imgproc::cornerHarris_caller(blockSize, (float)k, Dx, Dy, dst);\r
+}\r
+\r
+\r
 #endif /* !defined (HAVE_CUDA) */\r

diff --git a/tests/gpu/src/imgproc_gpu.cpp b/tests/gpu/src/imgproc_gpu.cpp
index 2c13334..12e1f32 100644 (file)
--- a/tests/gpu/src/imgproc_gpu.cpp
+++ b/tests/gpu/src/imgproc_gpu.cpp
@@ -603,6 +603,62 @@ void CV_GpuHistogramsTest::run( int )
     }\r
 }\r
 \r
+////////////////////////////////////////////////////////////////////////\r
+// Corner Harris feature detector\r
+\r
+struct CV_GpuCornerHarrisTest: CvTest \r
+{\r
+    CV_GpuCornerHarrisTest(): CvTest("GPU-CornerHarrisTest", "cornerHarris") {}\r
+\r
+    void run(int)\r
+    {\r
+        try\r
+        {\r
+            int rows = 1 + rand() % 300, cols = 1 + rand() % 300;\r
+            if (!compareToCpuTest(rows, cols, CV_32F, 1 + rand() % 5, -1)) return;\r
+            for (int i = 0; i < 3; ++i)\r
+            {\r
+                rows = 1 + rand() % 300; cols = 1 + rand() % 300;\r
+                if (!compareToCpuTest(rows, cols, CV_32F, 1 + rand() % 5, 1 + 2 * (rand() % 4))) return;\r
+            }\r
+        }\r
+        catch (const Exception& e)\r
+        {\r
+            if (!check_and_treat_gpu_exception(e, ts)) throw;\r
+            return;\r
+        }\r
+    }\r
+\r
+    bool compareToCpuTest(int rows, int cols, int depth, int blockSize, int apertureSize)\r
+    {\r
+        RNG rng;\r
+        cv::Mat src(rows, cols, depth);\r
+        if (depth == CV_32F) \r
+            rng.fill(src, RNG::UNIFORM, cv::Scalar(0), cv::Scalar(1));\r
+\r
+        double k = 0.1;\r
+        int borderType = BORDER_DEFAULT;\r
+\r
+        cv::Mat dst_gold;\r
+        cv::cornerHarris(src, dst_gold, blockSize, apertureSize, k, borderType); \r
+\r
+        cv::gpu::GpuMat dst;\r
+        cv::gpu::cornerHarris(cv::gpu::GpuMat(src), dst, blockSize, apertureSize, k);\r
+\r
+        cv::Mat dsth = dst;\r
+        for (int i = apertureSize + 2; i < dst.rows - apertureSize - 2; ++i)\r
+        {\r
+            for (int j = apertureSize + 2; j < dst.cols - apertureSize - 2; ++j)\r
+            {\r
+                float a = dst_gold.at<float>(i, j);\r
+                float b = dsth.at<float>(i, j);\r
+                if (fabs(a - b) > 1e-3f) return false;\r
+            }\r
+        }\r
+        return true;\r
+    }\r
+};\r
+\r
 /////////////////////////////////////////////////////////////////////////////\r
 /////////////////// tests registration  /////////////////////////////////////\r
 /////////////////////////////////////////////////////////////////////////////\r
@@ -620,3 +676,5 @@ CV_GpuNppImageIntegralTest CV_GpuNppImageIntegral_test;
 CV_GpuNppImageCannyTest CV_GpuNppImageCanny_test;\r
 CV_GpuCvtColorTest CV_GpuCvtColor_test;\r
 CV_GpuHistogramsTest CV_GpuHistograms_test;\r
+CV_GpuCornerHarrisTest CV_GpuCornerHarris_test;\r
+\r