Ascend NPU Image Processing

Goal

In this guide, you will gain insights into the thread safety of Ascend operators already in use, as well as discover how to effectively employ Ascend operators for image preprocessing and understand their usage limitations.

Preface

We provide a suite of common matrix operation operators that support the Ascend NPU within OpenCV. For user convenience, the new 'AscendMat' structure and its associated operators maintain compatibility with the 'Mat' interface in OpenCV. These operators encompass a wide range of frequently used functions, including arithmetic operations, image processing operations, and image color space conversion. All of these operators are implemented utilizing CANN(Compute Architecture of Neural Networks). The Ascend operator facilitates accelerated operations on the NPU by making use of CANN. This acceleration effect is particularly noticeable when working with larger images, such as those with dimensions like 2048x2048, 3840x2160, 7680x4320, etc.

Instructions on Thread Safety

Our stream function is implemented by invoking the CANN operators. In the same stream, tasks are executed sequentially, while across different streams, tasks are executed in parallel. The use of event mechanisms ensures synchronization of tasks between streams, please refer to the Stream Management documentation for details.

Example for Image Preprocessing

In this section, you will discover how to use Ascend operators for image preprocessing, including functions below:

• Add
• Rotate
• Flip

code

C++

// This file is part of OpenCV project.
// It is subject to the license terms in the LICENSE file found in the top-level directory
// of this distribution and at http://opencv.org/license.html.
 
#include <iostream>
#include <opencv2/imgcodecs.hpp>
#include <opencv2/cann.hpp>
#include <opencv2/cann_interface.hpp>
 
int main(int argc, char* argv[])
{
    cv::CommandLineParser parser(argc, argv,
                                 "{@input|puppy.png|path to input image}"
                                 "{@output|output.png|path to output image}"
                                 "{help||show help}");
    parser.about("This is a sample for image processing with Ascend NPU. \n");
    if (argc != 3 || parser.has("help"))
    {
        parser.printMessage();
        return 0;
    }
 
    std::string imagePath = parser.get<std::string>(0);
    std::string outputPath = parser.get<std::string>(1);
 
    // read input image and generate guass noise
    cv::Mat img = cv::imread(imagePath);
    // Generate gauss noise that will be added into the input image
    cv::Mat gaussNoise(img.rows, img.cols, img.type());
    cv::RNG rng;
    rng.fill(gaussNoise, cv::RNG::NORMAL, 0, 25);
 
    // setup cann
    cv::cann::initAcl();
    cv::cann::setDevice(0);
 
    cv::Mat output;
    // add gauss noise to the image
    cv::cann::add(img, gaussNoise, output);
    // rotate the image with a certain mode (0, 1 and 2, correspond to rotation of 90, 180 and 270
    // degrees clockwise respectively)
    cv::cann::rotate(output, output, 0);
    // flip the image with a certain mode (0, positive and negative number, correspond to flipping
    // around the x-axis, y-axis and both axes respectively)
    cv::cann::flip(output, output, 0);
 
    cv::imwrite(outputPath, output);
 
    cv::cann::resetDevice();
    cv::cann::finalizeAcl();
    return 0;
}

Python

# This file is part of OpenCV project.
# It is subject to the license terms in the LICENSE file found in the top-level directory
# of this distribution and at http://opencv.org/license.html.
 
import numpy as np
import cv2
import argparse
 
parser = argparse.ArgumentParser(description='This is a sample for image processing with Ascend NPU.')
parser.add_argument('image', help='path to input image')
parser.add_argument('output', help='path to output image')
args = parser.parse_args()
 
# read input image and generate guass noise
#! [input_noise]
img = cv2.imread(args.image)
# Generate gauss noise that will be added into the input image
gaussNoise = np.random.normal(0, 25,(img.shape[0], img.shape[1], img.shape[2])).astype(img.dtype)
#! [input_noise]
 
# setup cann
#! [setup]
cv2.cann.initAcl()
cv2.cann.setDevice(0)
#! [setup]
 
#! [image-process]
# add gauss noise to the image
output = cv2.cann.add(img, gaussNoise)
# rotate the image with a certain mode (0, 1 and 2, correspond to rotation of 90, 180
# and 270 degrees clockwise respectively)
output = cv2.cann.rotate(output, 0)
# flip the image with a certain mode (0, positive and negative number, correspond to flipping
# around the x-axis, y-axis and both axes respectively)
output = cv2.cann.flip(output, 0)
#! [image-process]
 
cv2.imwrite(args.output, output)
 
#! [tear-down-cann]
cv2.cann.finalizeAcl()
#! [tear-down-cann]

Explanation

Input Image

C++

    cv::Mat img = cv::imread(imagePath);
    // Generate gauss noise that will be added into the input image
    cv::Mat gaussNoise(img.rows, img.cols, img.type());
    cv::RNG rng;
    rng.fill(gaussNoise, cv::RNG::NORMAL, 0, 25);

Python

# Read the input image
img = cv2.imread("/path/to/img")
# Generate gauss noise that will be added into the input image
gaussNoise = np.random.normal(mean=0,sigma=25,(img.shape[0],img.shape[1],img.shape[2])).astype(img.dtype)

Setup CANN

C++

    cv::cann::initAcl();
    cv::cann::setDevice(0);

Python

cv2.cann.initAcl()
cv2.cann.setDevice(0)

Image Preprocessing Example

C++

    cv::Mat output;
    // add gauss noise to the image
    cv::cann::add(img, gaussNoise, output);
    // rotate the image with a certain mode (0, 1 and 2, correspond to rotation of 90, 180 and 270
    // degrees clockwise respectively)
    cv::cann::rotate(output, output, 0);
    // flip the image with a certain mode (0, positive and negative number, correspond to flipping
    // around the x-axis, y-axis and both axes respectively)
    cv::cann::flip(output, output, 0);

Python

# add gauss noise to the image
output = cv2.cann.add(img, gaussNoise)
# rotate the image with a certain mode (0, 1 and 2, correspond to rotation of 90, 180
# and 270 degrees clockwise respectively)
output = cv2.cann.rotate(output, 0)
# flip the image with a certain mode (0, positive and negative number, correspond to flipping
# around the x-axis, y-axis and both axes respectively)
output = cv2.cann.flip(output, 0)

Tear down CANN

C++

    cv::cann::resetDevice();
    cv::cann::finalizeAcl();

Python

cv2.cann.finalizeAcl()

Results

1. The original RGB input image with dimensions of (480, 640, 3):

puppy

1. After introducing Gaussian noise, we obtain the following result:

puppy_noisy

1. When applying the rotate operation with a rotation code of 0 (90 degrees clockwise), we obtain this result:

puppy_noisy_rotate

1. Upon applying the flip operation with a flip code of 0 (flipping around the x-axis), we achieve the final result:

puppy_processed_normalized