Note

This is the documentation for the latest development branch and may refer to features that are not available in released versions. If you are looking for the documentation for a specific release, use the drop-down menu on the left and select the desired version.

OpenCV Example

OpenCV Overview

OpenCV (Open Source Computer Vision Library) is a cross-platform computer-vision library originally initiated by Intel and released under the BSD license. It can be used freely in both commercial and research scenarios. OpenCV is widely used to build real-time image processing, computer vision, and pattern-recognition applications.

K230 RTOS SDK provides an optimized OpenCV acceleration library for K230. Compared with the original OpenCV version, it significantly reduces the runtime of many OpenCV operators. For example:

Operator	K230 + original OpenCV	K230 + optimized OpenCV
integral image	34.5 ms	7.7 ms
warp affine	391.1 ms	34.7 ms

Notes:

• The above timings were measured on the K230 big core at 1.6 GHz.

• For the integral-image test, the input image is a 1280 x 1080 grayscale image and the integral image uses 32-bit float.

• For the warp-affine test, the input image is a 1280 x 1080 grayscale image, rotated 15 degrees clockwise and scaled by 0.6, with a 1280 x 1080 grayscale destination image.

The SDK already includes a prebuilt optimized static OpenCV library under:

src/rtsmart/libs/opencv/

Users can directly link against this library when building their own applications.

Build Examples

This section explains how to build executables with the OpenCV static library already provided in the SDK. The SDK provides five example programs under:

src/rtsmart/examples/3rd-party/opencv

Code Structure

Representative directory structure:

opencv_calculate_hist
opencv_detect_features2d
opencv_find_contours
opencv_grayscale_binarize
opencv_obj_detect
CMakeLists.txt
build_app.sh
cmake/
utils/

The utils directory contains all required input images and related data.

Firmware Build

If you want the examples to be built into the firmware, run make menuconfig from the SDK root and enable:

RT-Smart UserSpace Examples Configuration
-> Enable build 3rd-party examples
-> Enable Build OpenCV Sample Programs

The built examples are placed under sdcard/app/examples/3rd-party/opencv.

Example Build

If you only want to build the OpenCV examples, enter:

src/rtsmart/examples/3rd-party/opencv

and run:

./build_app.sh

After a successful build, the k230_bin directory contains the compiled elf files and the required test files.

Run Examples

Note

All required input images are provided under:

src/rtsmart/examples/3rd-party/opencv/utils

opencv_calculate_hist

This sample reads an image and calculates the histogram of each color channel (B, G, R).

./opencv_calculate_hist.elf

opencv_detect_features2d

This sample reads an image and uses the FAST detector to detect image feature points.

./opencv_detect_features2d.elf

opencv_find_contours

This sample reads an image, detects contours, and draws the contour result into a new image.

./opencv_find_contours.elf

opencv_grayscale_binarize

This sample reads a color image, converts it to grayscale, then performs binarization and saves the processed result.

./opencv_grayscale_binarize.elf

opencv_obj_detect

This sample uses cascade classifiers to detect faces and eyes, then draws the detection result on the original image.

./opencv_obj_detect.elf