
Set up python with pyenv on mac, create a dedicated OpenCV development virtual environment, then install numpy and opencv, verify by importing cv2, and configure PyCharm to use the env.
This is a gentle introduction to what to expect in this the course and what you can expect to learn.
This is an introduction to OpenCV
Learn how to read an image with OpenCV, display it with imshow, and understand that an image is a multidimensional array of pixels with blue, green, and red channels.
Convert a color image to grayscale by collapsing the three color channels into a single 8-bit value (0-255), and compare the original and grayscale results.
Build and understand images by coding a 100x100 blank array, create a random color image with three channels, and practice a binary image (0 or 255).
Create a random binary image by filling a 100x100 grid with 0 or 255 values, then save it to a file using imwrite.
Encode a secret message in the Mona Lisa image by converting characters to pixel values and updating a row, then save and prepare for extraction.
Decode a hidden message embedded in the Mona Lisa image using Python, converting numeric pixel data from the thousandth row back into characters to reveal the text.
Learn how to treat a video as a sequence of frames, access them via a stream, and perform frame-level manipulations, including overlaying elapsed time to create a new video.
OpenCV pro teaches capturing live video from a camera using index 0, resolves a black screen by pausing one millisecond, and runs a live feed.
Overlay a live timestamp on video frames with OpenCV text rendering, handle BGR colors, compute elapsed time, and prepare frames for saving to a new video file.
Learn to capture and save manipulated video by creating a video writer, configuring frame size and fps from the camera, and writing frames to an MJPEG file.
Count Eminem candies by building image pipeline that converts to grayscale, reduces noise with blur, creates a binary image with white foreground and black background, and analyzes contours to count.
Convert the image to grayscale, apply gaussian blur to reduce noise, then threshold to produce a binary image for counting candies.
Learn adaptive thresholding with OpenCV to transform blurred grayscale images into precise binary images by computing local thresholds from nearby pixels, enabling robust contour-based counting.
Count and validate image contours with OpenCV using external mode and simple retrieval, thresholding, and area-based noise filtering. Plan to draw the contours on the image in the next video.
Count contours in OpenCV fundamentals using a binary image created with adaptive threshold. Filter by area and highlight contours on the color image, then display the total count.
Record a live video feed and run the Eminem candy counting algorithm on frame-by-frame data. Apply thresholding, adaptive threshold, and contour detection to count candies in real time.
Refactor the candy counting into a reusable function that takes an image and returns the candy count via contour detection and area thresholding, applying it to video frames show counts.
Use OpenCV to locate the queen on a carrom board by grayscale conversion, color cues, blur, and in range thresholding to form a binary image, then detect contours.
Learn to detect a color object using the Lab color space and inRange thresholding, converting from BGR to Lab, creating a binary image, and extracting contours to locate the queen.
learn to detect the queen by converting to Lab, creating a binary image, finding contours, selecting the largest by area, and drawing it when the area exceeds a threshold.
Compute the queen’s x and y coordinates from a contour using a minimum enclosing circle and a bounding rectangle, then highlight the position with a circle.
Master thresholding with Otsu, masks, and the lab color scheme to create binary images for detecting employee t-shirts in video or images, with erosion and contour analysis to reduce noise.
preprocess a photo by grayscale conversion and Gaussian blur, apply adaptive and simple thresholding to create a binary image, then detect the t-shirt contour for analysis.
Create a mask from contours by drawing on a blank image and filling the target region, then apply a bitwise and to extract the desired area such as a t-shirt.
Learn to compute and plot color histograms with calcHist and matplotlib, using a mask to analyze the t-shirt and convert to Lab color space for robust detection.
Develop an employee finder that uses t-shirt color properties and contour analysis in OpenCV to detect, annotate, and return images showing whether an employee is present.
Analyze live video with OpenCV by using an employee finder to detect employees in each frame and highlight them using contour-based detection from a live iPhone video stream.
Build an image processing pipeline in OpenCV to count eggs inside a box, using contour analysis, region of interest masks, erosion, and circle-like heuristics.
Apply grayscale conversion, Gaussian blur, and thresholding to create a binary image, then detect external contours and approximate the largest contour to isolate the box.
Create a binary mask of the target region, draw a white rectangle on a black mask, then apply bitwise_and to extract the region of interest for contour analysis.
Learn how to select contours by size in OpenCV, convert to grayscale, blur, threshold to binary, remove small contours, and analyze contour areas to identify egg-like shapes.
OpenCV Pro lecture demonstrates counting eggs by filtering 17 contours in a box, applying circle detection and a circle-area heuristic to identify eggs and display the count.
Explore how deep learning integrates with image analysis using OpenCSV to extract contours inside a box and classify objects like eggs versus other items, leveraging feature extraction before machine learning.
This course covers the fundamental of OpenCV and it is designed to be project oriented. Instead of going over command after command available in OpenCV, we will be tackling various computer vision problems. And then while working through each vision problem, we will learn new openCV commands and apply them to solve the vision problem at hand.
Below are 7 vision problems we will work through. The vision problems in the beginning are rather simple where you will encounter many new concepts and commands. As we progress through, the problems will get harder and they will build on top of the knowledge you would have gained in the previous problems.
Encode/Decode Secret Messages in the Mona Lisa Picture
Adding time Elapsed to an existing video, like in those director's cut movies
Determine how many pieces of candy are left
Count Candy pieces off Camera Video
Finding the queen in the game of carrom board
Determine if an employee wearing company t-shirt is present on the floor via camera
Count the number of eggs in the box
Here are the most common concepts your will learn and apply in this course:
Image & Video Data Structure
Image Pre-processing (Erode, Dilate etc.)
Contour Detection
Region of Interest Detection
Color Space Manipulation
Image Processing Pipeline