Merge pull request #26621 from CodeLinaro:apreetam_2ndPost

FastCV-based HAL for OpenCV acceleration 2ndpost-3 #26621

### Detailed description:

- Add cv_hal_canny for Canny API

Requires binary from [opencv/opencv_3rdparty#90](https://github.com/opencv/opencv_3rdparty/pull/90) 
Depends on: [opencv/opencv#26617](https://github.com/opencv/opencv/pull/26617)
Depends on: [opencv/opencv#26619](https://github.com/opencv/opencv/pull/26619) 

### Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [ ] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
      Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake

3rdparty/fastcv/include/fastcv_hal_imgproc.hpp

@@ -26,6 +26,8 @@
 #define cv_hal_cvtBGRtoHSV          fastcv_hal_cvtBGRtoHSV
 #undef  cv_hal_cvtBGRtoYUVApprox
 #define cv_hal_cvtBGRtoYUVApprox    fastcv_hal_cvtBGRtoYUVApprox
+#undef  cv_hal_canny
+#define cv_hal_canny                fastcv_hal_canny
 ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
 /// @brief Calculate medianBlur filter
 /// @param src_data Source image data
@@ -235,4 +237,32 @@ int fastcv_hal_cvtBGRtoYUVApprox(
     int             scn,
     bool            swapBlue,
     bool            isCbCr);
+
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+/// @brief Canny edge detector
+/// @param src_data Source image data
+/// @param src_step Source image step
+/// @param dst_data Destination image data
+/// @param dst_step Destination image step
+/// @param width Source image width
+/// @param height Source image height
+/// @param cn Number of channels
+/// @param lowThreshold low hresholds value
+/// @param highThreshold high thresholds value
+/// @param ksize Kernel size for Sobel operator.
+/// @param L2gradient Flag, indicating use of L2 or L1 norma.
+////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+int fastcv_hal_canny(
+    const uchar* 	src_data,
+    size_t          src_step,
+    uchar* 			dst_data,
+    size_t          dst_step,
+    int             width,
+    int             height,
+    int             cn,
+    double          lowThreshold,
+    double          highThreshold,
+    int             ksize,
+    bool            L2gradient);
+
 #endif

3rdparty/fastcv/src/fastcv_hal_imgproc.cpp

@@ -992,4 +992,59 @@ int fastcv_hal_cvtBGRtoYUVApprox(
 
     fcvStatus status = FASTCV_SUCCESS;
     CV_HAL_RETURN(status, hal_BGRtoYUVApprox);
+}
+
+int fastcv_hal_canny(
+    const uchar*    src_data,
+    size_t          src_step,
+    uchar*          dst_data,
+    size_t          dst_step,
+    int             width,
+    int             height,
+    int             cn,
+    double          lowThreshold,
+    double          highThreshold,
+    int             ksize,
+    bool            L2gradient)
+{
+    int numThreads = cv::getNumThreads();
+
+    if(numThreads!=1)
+        CV_HAL_RETURN_NOT_IMPLEMENTED("API performs optimally in single-threaded mode");
+
+    if (cn != 1)
+        CV_HAL_RETURN_NOT_IMPLEMENTED("Multi-channel input is not supported");
+
+    if (lowThreshold > highThreshold)
+        CV_HAL_RETURN_NOT_IMPLEMENTED("lowThreshold is greater then highThreshold");
+	
+	const double epsilon = 1e-9;
+	
+	if (std::abs(lowThreshold - std::round(lowThreshold)) > epsilon || std::abs(highThreshold - std::round(highThreshold)) > epsilon)
+        CV_HAL_RETURN_NOT_IMPLEMENTED("threshold with decimal values not supported");
+
+    INITIALIZATION_CHECK;
+
+    fcvStatus               status;
+    fcvNormType             norm;
+
+    if (L2gradient == 1)
+        norm = fcvNormType::FASTCV_NORM_L2;
+    else
+        norm = fcvNormType::FASTCV_NORM_L1;
+
+    if ((ksize == 3) && (width > 2) && (height > 2) && (src_step >= (size_t)width) && (dst_step >= (size_t)width))
+    {
+        int16_t* gx = (int16_t*)fcvMemAlloc(width * height * sizeof(int16_t), 16);
+        int16_t* gy = (int16_t*)fcvMemAlloc(width * height * sizeof(int16_t), 16);
+        uint32_t gstride = 2 * width;
+        status = fcvFilterCannyu8(src_data, width, height, src_step, ksize, static_cast<int>(std::round(lowThreshold)), static_cast<int>(std::round(highThreshold)), norm, dst_data, dst_step, gx, gy, gstride);
+        fcvMemFree(gx);
+        fcvMemFree(gy);
+    }
+    else
+    {
+        CV_HAL_RETURN_NOT_IMPLEMENTED(cv::format("Ksize:%d is not supported", ksize));
+    }
+    CV_HAL_RETURN(status, hal_canny);
 }