How does a camera capture an image?
One-line takeaway
BLUF: A lens bends incoming light so rays from each scene point converge onto one point on a sensor. A shutter sets how long that lasts, an aperture sets how much light gets through, and the sensor records brightness at millions of separate positions.
Once you see a photograph as a controlled trade between light, time and depth, exposure settings stop being arbitrary numbers.
Getting every ray back to one point
Light leaves every point of a scene in all directions. A bare sensor would receive light from everywhere at once and record grey mush. A pinhole fixes this by admitting only one ray per point, but it is dim. A lens does better: its curved glass refracts many rays from a single scene point back together at a single image point, so the picture is both sharp and bright. Behind the lens sits the sensor, a grid of millions of photosites. Each one collects photons during the exposure and outputs a voltage proportional to how many arrived. A colour filter array, usually a Bayer pattern of red, green and blue, sits on top, so each site measures one colour and software interpolates the rest.
Light, time, and a cone of directions
The deeper principle is that a photograph is an integral of light over a chosen slice of time and a chosen cone of directions. Three controls set that slice. Shutter speed sets duration, which fixes how much motion blurs. Aperture sets the diameter of the opening, which changes brightness and simultaneously changes depth of field, because a narrow opening accepts a narrower cone of rays and tolerates more focus error. Sensitivity, still labelled ISO, amplifies the signal afterward and amplifies noise with it. Any two of these can be traded against the third for the same exposure but a different-looking picture. Focus is just moving the lens so the convergence plane lands on the sensor.
From phone sensors to space telescopes
The same optics runs from phone cameras to space telescopes. Phones compensate for tiny lenses and tiny sensors with computation: they capture bursts, align and stack them to cut noise, merge different exposures for dynamic range, and estimate depth to fake shallow focus. Scientific and medical imaging uses the same sensors tuned for wavelengths the eye cannot see. Digital sensors come in two main families, CCD and CMOS, with CMOS now dominant because it is cheap and integrates readout on the chip. Rolling shutters, which read the sensor line by line, are why fast-moving objects sometimes appear skewed in video from ordinary cameras.
Megapixels and other red herrings
Common misconceptions start with resolution. Myth: more megapixels means a better photograph. Reality: sensor size, lens quality, and light gathered per pixel usually matter more, and cramming pixels onto a small sensor can increase noise. Myth: zooming and moving closer are equivalent. Reality: focal length changes perspective and how compressed the background looks, so they produce different images. Myth: the camera captures what your eye sees. Reality: the eye has far greater dynamic range and continuously adapts, which is why bright skies blow out in photographs. Myth: raw files are unprocessed truth. Reality: they still need demosaicing, white balance and tone mapping; they simply defer those decisions instead of baking them in.