It's interesting how the vertical lines are slightly curved. From lecture, each pixel (sensor position) is represented by a fixed position $x$ over all positions $u$ in the Cartesian ray space. Each pixel is the integration along the vertical line. Could the curves be due to the curvature or imperfection of the lens, which affects position of $u$?

ellenluo

I'm confused about what x and u represent in this chart. I'm also interested in why the line is curved.

hershg

Could the curvature have something to do with the actual hardware limitations/how the sensor refreshes and captures data at each pixel (or the 4D analog for this case)? kinda similar to the 'rolling' effect from camera captures we saw a few lectures ago?

hershg

From a slide at the end in the extras section, it appears that curves are due to lens aberrations

It's interesting how the vertical lines are slightly curved. From lecture, each pixel (sensor position) is represented by a fixed position $x$ over all positions $u$ in the Cartesian ray space. Each pixel is the integration along the vertical line. Could the curves be due to the curvature or imperfection of the lens, which affects position of $u$?

I'm confused about what x and u represent in this chart. I'm also interested in why the line is curved.

Could the curvature have something to do with the actual hardware limitations/how the sensor refreshes and captures data at each pixel (or the 4D analog for this case)? kinda similar to the 'rolling' effect from camera captures we saw a few lectures ago?

From a slide at the end in the extras section, it appears that curves are due to lens aberrations