The Human Eye: Shutters & Exposure

You spin your robot. You take a picture. It’s a blurry mess. Your code crashes because it can’t find the target. You say “The camera is bad.” No. The settings are bad. To fix it, you have to understand how your own eyes work.

The Iris vs. The Aperture

Walk into a dark room. Your pupil (the black dot) gets huge. It’s trying to let in as much light as possible. Walk into the sun. Your pupil becomes a tiny pinprick.

• Too much light: The image is “blown out” (White screen).
• Too little light: The image is noisy and dark.
• Robotics: We manually lock the Exposure (Brightness) of our webcams. If we leave it on “Auto,” the flashy stadium lights confuse the camera, causing it to darken the image until we can’t see the game pieces.

The Shutter Speed (Motion Blur)

This is the big one. Your eye doesn’t have a shutter, but your brain processes images at a certain frame rate. A Camera sensor “collects” light for a specific time window.

• Long Exposure (1/30th sec): The shutter stays open for a while. It gathers lots of light. The image is bright. BUT, if the robot moves during that time, the image smears.
• Short Exposure (1/1000th sec): The shutter snaps open and shut instantly. It freezes time. The image is crisp. BUT, it’s dark because not enough light got in.

The Engineer’s Dilemma

In robotics, we need Sharp Images (to read AprilTags or detect rings). But competition fields are often dimly lit gyms. If we lower the Exposure (to stop blur), the image gets too dark to see.

• The Solution: We buy “Global Shutter” cameras (like the Arducam OV2311) sensitive to low light, or we add LED Rings (Flashlights) to our robot. we literally bring our own sun so we can run the shutter spread faster.

Conclusion

Computer Vision isn’t just code; it’s photography. If you treat your webcam like a sensor and tune its physics (Exposure, Gain, White Balance), your code will work 10x better. Don’t write more code to fix a bad image. Fix the image physics first.