HAW: Lesson - How Computer Vision Works
How Computer Vision Works
Check out the sensory signals on the graphic below:
Dark Blue: Ultrasound. Directly surrounds the car. Adaptive cruise control.
Blue: Camera. Rear: Park assistance, surround view. Sides: Surround view. Front: Lane departure warning, traffic signal recognition.
Green: Long-Range RADAR. Rear: Rear collision warning. Rear side: Blind spot detection. Front: Cross traffic alert.
Orange: LIDAR. Emergency braking, pedestrian detection, collision avoidance.
Yellow: Wireless communication
In this exciting time of change, our lives and future are being shaped by diverse individuals and collaborative teams within our global community. Below are examples of some of those individuals, their contributions, and the impact of their contributions. As you progress throughout this lesson, think about how our modern-day advancements were influenced by the innovators below. For more information about their lives, careers, innovations, and advancements select the designated source link.
Let's take a look at a second image, which highlights RADAR:
- Green: Driver Monitor RADAR - 3D positioning, driver alertness, driver awareness
- Orange: Short Range/Medium Range RADAR - park assist, cross-traffic alert, junction assist
- Purple: Medium Range RADAR - blind side detection
- Blue: Long Range RADAR - adaptive cruise control, automatic emergency braking, forward collision warning
Delivery Vehicle of the Future?
The following video features an autonomous vehicle called the Nuro R2 (nuro.ai). It is used for food delivery. Look carefully at the vehicle as you watch the video and afterwards, review the list of technological features of this autonomous vehicle.
Vehicle Features:
- Narrow width
- A more narrow vehicle means it takes up less road space, which means it will be safer for other vehicles and pedestrians.
- Pedestrian-Protecting Front End
- A special front panel absorbs energy and ultrasonics to protect pedestrians.
- Additionally, there is a sound-generating device in the front bumper to increase pedestrian awareness of the vehicle's proximity
- Multiple cameras housed on top of the vehicle
- 360° view using overlapping standard cameras
- A front viewing thermal camera
- Lidar
- Both short and long-range radar (also located in the front grill)
- Curbside Delivery Doors
- Customers gain access to their delivery on the curbside, avoiding having to step into traffic.
- A touch screen panel has interfaces for both customers and law enforcement
- Additional Features
- Redundant braking and control
- Automotive lighting and signaling
- Specifications:
- Width: 1.10m
- Length: 2.74m
- Height: 1.86m
- Max speed: 25mph
- Battery size: 31kWh
- Charge speed: L2, 6.6kWh/hr
- Gross Vehicle Weight: 1150kg
- Payload: 190kg
- Carrying Capacity: 22.38ft3
All information provided by the creator of the Nuro R2, nuro.ai.
How Computer Vision Works
So, all of that is really impressive. But let's put it all together to fully understand this question: just how can a computer see?
How can an autonomous vehicle tell the difference between a tree, a person, or a stop sign? How do face filters work? How does facial recognition work? Watch this short, but informative video of how patterns, edges, and neural networks all work together when utilizing computer vision, all courtesy of Code.org:
Practice Activity
What sorts of things do you think autonomous vehicles look for? Let's do a quick practice activity to review.
[CC BY-NC-SA 4.0
Links to an external site.] UNLESS OTHERWISE NOTED | IMAGES: LICENSED AND USED ACCORDING TO TERMS OF SUBSCRIPTION - INTENDED ONLY FOR USE WITHIN LESSON.
Image 1: metamorworks/Shutterstock.com. Image used under license from Shutterstock.com and may not be repurposed.
Image 2: Image provided by https://nxp.com