Learn Computer Vision with OpenCV and Python

Explore OpenCV thresholding with four methods—global, adaptive mean, adaptive gaussian, and otsu—and learn how binary conversion and neighborhood area shape edges for computer vision preprocessing.

Explore convolution with kernels and filter2D, applying averaging, Gaussian, and median blur, then Sobel and Laplacian edge detection for cnn feature learning.

Explore morphological transformations in images with kernels, eroding boundaries and applying deflation to foreground shapes. Observe opening and closing operations that refine images and fill holes using kernels and parameters.

Learn to implement corner detection in Python with OpenCV, converting images to grayscale and applying Harris corner detection, identifying corner points and visualizing them in red.

Learn how to detect key points and describe them with descriptors, then match features across images to identify corresponding regions using OpenCV in Python.

Explore contour properties with OpenCV and Python, computing area, arc length, and bounding rectangles for contours, and use thresholding and morphology to differentiate shapes by solidity, aspect ratio, and extent.

Detect blobs in grayscale images with the simple blob detector in OpenCV, adjust area, circularity, and inertia filters to identify circles and holes in example shapes.

This lecture demonstrates circle detection in a grayscale image using a circle-detection function, draws circle and its center, and explains parameters such as input image, method, edge-detection, and minimum distance.

Learn to detect lines in images using a transform on grayscale edges, and tune parameters such as distance resolution, angle resolution, minimum line length, and maximum gap.

Learn how to apply watershed segmentation in OpenCV with Python to separate touching coins, using grayscale conversion, thresholding, distance transform, local maxima, and contour labeling.

Apply OpenCV background subtraction to count people by separating the foreground, extracting contours, tracking contour centers across frames, and counting those crossing a horizontal reference line.

Filter objects by color with HSV space in OpenCV and Python, applying red, green, and blue thresholds to detect contours and track the object's center from a webcam or video.

Explore moving object tracking with OpenCV and Python, using background subtraction, contour detection, and blob-based tracking to analyze foreground motion and output video results.

Explore haarcascade face and eye detection in OpenCV by training with positive and negative images, converting frames to grayscale, and applying multi-scale detection.

Explore histogram of oriented gradients (HOG) features and an OpenCV SVM-based pedestrian detector, learning to extract HOG descriptors and apply non-maximum suppression to detect pedestrians.

Explore how to build a logo search app by extracting key points and descriptors, comparing logos in a dataset, and scoring similarity to identify the best match.

Learn how to build a simple image dataset by automatically downloading Google images with Python, organizing files, and preparing data for labeling and bird-detection modeling.

Label images with LabelImg to create bounding boxes, building a bird detection dataset with xml annotations for training and testing a deep learning model and not bird regions.

Train your model with a simple dataset for bird detection using keras to build a sequential cnn, apply data augmentation, and save the trained model as birds.model.

Test a trained bird-detection model with keras by selecting an ROI, resizing and normalizing the patch, loading the model, and predicting whether the patch is a bird.