OpenCV Python Development: Detecting Objects

Name: OpenCV Python Development: Detecting Objects
Author: Tom Rosenwinkel

Learn how to detect objects and describe their features in OpenCV

Preview the first lesson free — get full access to all 11 lessons.

Course: On-Demand

Hello, my name is Tom Rosenwinkle, and in this series of lessons, we're going to learn several methods that are applied toward object detection in OpenCV Python development. Before we start learning those methods, I would like to show the interface that we'll be using to demonstrate these methods. I've created an interface between OpenCV and a package called Streamlit. For our image processing that we'll do for this course, we will work in a file that I have named OpenCV. We will work in a file that I have named OCV process, and creating a function called process image. Right now it is empty. The way that we will use that in the interface is that I've set up a video capture object configured to my webcam, and using the Streamlit interface, I'm creating two columns on the web app that we'll launch. In the left column, I'm creating an area to display the frame from the camera as a source, and in the right column, I'm doing the same thing, creating an area to present the updated image that is processed, because the camera will continuously take frames, new frames, and I want to read them as long as the camera is taking them in an infinite loop, I'm creating a button that can also be used to break the loop. So if the Streamlit button that is rendered will be clicked, it will set my break loop variable to true, otherwise it is false. So we can continuously read frames from the webcam as long as the frame was read successfully, and as long as the break loop button was not clicked, then the loop will proceed. I'm doing some basic processing on the frame to interface between OpenCV and Streamlit. Streamlit, like many interfaces, expects images to be in a color scheme of red, green, blue, but we remember that OpenCV's color scheme is blue, green, red. So I convert between the color schemes using OpenCV's, are using OpenCV's BGR to RGB flag, color flag. And also, I flip the columns of the image so that as it is displayed and I move an object between left and right, it will display it between left and right as I expect. So then that frame area for the source from the webcam is updated with that version of the frame, and then we copy the frame to process it and update it in the destination side or the out area that was blocked off at the beginning of the script. So to launch the web app and to run this interface, we'll use the Streamlit executable with the run command and the name of our source code file. I've saved it in my project directory and a subdirectory called SRC. I've called it home.py. So when I launch the Streamlit app with this command, the script will be run and we can see the room where I'm recording these videos. You can see the ceiling fan moving. You can see I can reach my hand up into the view of the web camera and move to the left and to the right. You also see that the two images that are displayed are equivalent because we've not done any processing in our function file. When I click the stop capture button that I created, that also will break the infinite loop. So Streamlit will rerun the script as often as I either click rerun or even if I make an update to the source code or otherwise every time I interact with the web app, it will re-execute the script from top to bottom. In the next lesson, we'll use this interface and we'll see our first methods toward object detection. We will see examples of binary images and bitwise operations on those images. Thank you for watching.

Intermediate Provider : Tom Rosenwinkel 11 Lessons · 53m in English

Course Description

How can you identify and locate objects in an image or video frame? How can you describe their shape characteristics and match these objects? OpenCV provides methods for several techniques that can be used separately or together to detect objects by their shapes, color ranges, and contour measurements. Learning how to detect objects of interest in images and video frames is an essential skill in many advanced computer vision and machine learning projects. The topics discussed in these lessons will provide a foundation for future work in computational image enhancement or computer vision AI applications.

In these lessons, you will learn how to detect objects in images and video frames using OpenCV. We will explore binary and grayscale thresholding techniques, edge detection, and contour analysis. By the end of these lessons, you will be able to build a skin detector and describe the objects detected in an image.

What You'll Learn

Detect objects in images and video frames using OpenCV by their shapes, color ranges, and contour measurements
Apply binary and grayscale thresholding techniques, including threshold types and adaptive thresholding, for object detection
Perform edge detection, including Canny edge detection, to identify objects of interest
Analyze contours through measures, curvatures, and bounding contours to describe detected objects
Build a skin detector and describe the objects detected in an image
Explore best practices for detecting objects and matching them in OpenCV

Key Takeaways

OpenCV provides methods for several techniques—detecting objects by shapes, color ranges, and contour measurements—that can be used separately or together.
Thresholding techniques are important for object detection in OpenCV, covering binary and grayscale thresholding, threshold types, and adaptive thresholding.
Edge detection and contour analysis are core skills for identifying and describing objects in images and video frames.
Learning to detect objects of interest is an essential skill for advanced computer vision and machine learning projects.
These lessons provide a foundation for future work in computational image enhancement or computer vision AI applications.

Frequently Asked Questions

What will I be able to do by the end of this course?

By the end of these lessons, you will be able to detect objects in images and video frames using OpenCV, build a skin detector, and describe the objects detected in an image.

What object detection techniques does this course cover?

The course explores binary and grayscale thresholding techniques, edge detection (including Canny edge detection), and contour analysis. It also covers detecting objects by their shapes, color ranges, and contour measurements.

What skills will I gain from this course?

You will gain skills in Computer Vision, Feature Detection, Image Analysis, Object Detection, Object Recognition, and OpenCV.

Who is this course suited for?

It is suited for those building advanced computer vision and machine learning projects, as detecting objects of interest in images and video frames is described as an essential skill that provides a foundation for future work in computational image enhancement or computer vision AI applications.

What topics are taught in the lessons?

Lessons include Overview of Object Detection, Binary Images and Bitwise Operations, Threshold Types and Adaptive Thresholding, Skin Detection, Edge Detection, Canny Edge Detection, Contours, Measures of Contours, Curvatures of Contours, Bounding Contours, and Object Detection.