Introduction to Computer Vision

---

---

Computer vision is one of the most exciting fields in artificial intelligence, enabling computers and robots to interpret and understand visual information from the world around them. From detecting objects and recognizing faces to estimating depth and enabling autonomous navigation, computer vision bridges the gap between raw camera data and intelligent decision-making.

This webinar introduces participants to the fundamentals of computer vision through a practical, hands-on-oriented exploration of how vision systems work. We will begin by examining how cameras capture light and convert it into digital image data, including the role of image sensors, pixels, color channels, and image frames. We will then explore how software tools such as OpenCV can be used to process, manipulate, and extract useful information from these images.

Participants will be introduced to key computer vision techniques, including color-based detection, image features, Histograms of Oriented Gradients (HOG), and basic facial feature analysis. We will also explore the differences between monocular and stereo vision, discussing how a single camera can be used to identify visual features and how two cameras can help estimate depth and understand spatial relationships.

The session will then progress from traditional computer vision techniques to modern deep-learning-based approaches, introducing object detection models such as YOLO and RF-DETR and explaining how these models identify, classify, and locate objects within images.

Finally, we will connect these concepts to real-world robotics through our experience with the Vulcan competition, demonstrating how camera-based perception can support autonomous robotic systems. We will explore how a robot can use visual information, including object recognition and color detection, to interpret its environment and make navigation decisions, even when relying on a camera as its primary perception sensor.

Whether you are interested in artificial intelligence, robotics, autonomous systems, or software development, this webinar will provide a foundation for understanding computer vision and the tools and techniques used to build practical vision-based applications.

No prior experience in computer vision is required. A basic familiarity with programming is helpful but not essential.

Meeting Link: https://meet.google.com/dmg-fypi-wqc

---

---

---

---

AGENDA: INTRODUCTION TO COMPUTER VISION

1. Welcome and Introduction to Computer Vision

• What is computer vision?

• Computer vision vs. image processing vs. artificial intelligence

• Real-world applications of computer vision in robotics, healthcare, security, and autonomous systems

2. How Cameras See the World

• How light is captured by image sensors and converted into digital images

• Understanding pixels, resolution, RGB/BGR color channels, and image frames

• Camera data pipelines and frame rates

• How raw camera data becomes useful information

3. OpenCV and Fundamentals of Image Processing

• Introduction to OpenCV

• Reading, displaying, and processing images and video frames

• Image resizing, cropping, filtering, thresholding, and drawing

• Understanding color spaces

• Practical demonstration: detecting colored objects using OpenCV

4. Traditional Computer Vision Techniques

• Image features, edges, shapes, and contours

• Color detection and image segmentation

• Introduction to Histograms of Oriented Gradients (HOG)

• Basic facial feature analysis and face detection

• Traditional computer vision vs. machine-learning-based approaches

5. Monocular vs. Stereo Vision

• Understanding single-camera (monocular) vision

• How stereo cameras estimate depth

• Camera calibration, disparity, and triangulation

• Applications in robotics, distance estimation, and obstacle awareness

6. Modern Computer Vision: YOLO and RF-DETR

• Introduction to deep-learning-based computer vision

• Understanding object detection, classification, confidence scores, and bounding boxes

• Introduction to YOLO and RF-DETR object detection models

• Comparing traditional computer vision techniques with deep-learning approaches

• Integrating object detection models into camera-based applications and robotic systems

7. Computer Vision in Action: Lessons from the Vulcan Competition 

• Overview of our experience implementing autonomous behavior using a camera

• The camera-to-action pipeline: image capture, preprocessing, perception, decision-making, and control

• Applying object recognition and color detection to robotic autonomy

• Challenges involving lighting, perspective, processing latency, and detection reliability

• Lessons learned from building a camera-based autonomous system

8. Questions, Discussion, and Closing 

• Summary of key concepts

• Audience questions and discussion

• Recommended tools, resources, and next steps for learning computer vision

---

Join us as we explore the technology behind intelligent vision systems

Come ready to explore, learn, and discover how you can begin building your own computer vision applications.

Hosted by: IEEE JKUAT Computer Society Student Branch Chapter

Meeting Link: https://meet.google.com/dmg-fypi-wqc

---