OpenCV (Open Source Computer Vision Library) is an open-source computer vision and machine learning software library. It provides a wide range of algorithms and functions that enable developers to perform various image and video processing tasks. OpenCV is written in C++ and has interfaces for various programming languages, including Python.
# Performing basic image processing operations like resizing, cropping, rotating, and flipping images. # Cross platform Supports Many Language such as C++, Java, Python # Image, Video, Camera related Computer Vision, exist in Open CV # Computer Vision Applications # Applying various image filters and enhancements, such as blurring, sharpening, and adjusting brightness/contrast ( HPF, LPF). # Detecting and recognizing objects, faces, and text in images and videos. # Extracting features from images, such as edges, corners, and keypoints ( Feature Extraction). # Performing image segmentation and finding contours. # Calibrating cameras and working with camera parameters. # Pattern recognition - color, pixels value - object can be recognize # Photogrammetry - measurements - weights, height, volume # Open CV AR, VR
# 3 types of image - Read by Open CV # Color image, BW image # Loads image RGB alpha channel # Image Filtering # Find out the content Features # Filter, High Pass Filter (HPF) , Low Pass Filter (LPF) # Object can be identified # Low pass filter is noise such as you can see when click photos during night
# You have to create matrix and apply convolution # Convolution - Feature Extraction - Dimensional Reduction # Also called Kernel ( also LPF, HPF ) # After then It is called Convoluted image # Pixels values change so we can detect by our eyes # Eg. Black Hole image is created using pixels since it can't be captured and seen # 3 by 3 matrix, 5 by 5 which can be larger, 2 by 2 can be smaller # Normalize
Harr-cascade Based on opencv principle another application is made which is called harr-cascade # Mainly focused on face # Viola-Jones algorithm two scientist # Later it is widely used for object detection # If you want light application then you can use harr-cascade, but for accurate result deep learning is necessary # In present, Feature loss problem is a major problem # If you want light application then you can still use this # For accurate result deep learning is necessity # Harr-cascade operates on Black n White image # Lighting issue on phot could not recognize the person # Could not recognize the black people
# Code import cv2 image = cv2.imread('/content/salon_rai.jpg') image --> array([[[194, 194, 182], [195, 195, 183], [194, 194, 182], ..., [ 57, 46, 38], [ 12, 6, 1], [ 6, 1, 0]], [[197, 197, 185], [198, 198, 186], [197, 197, 185], Color images consist of 3 dimension