Fractional sensor names look like nonsense left over from 1950s television tubes - a 1/1.7" sensor is obviously not 0.59 inches of anything. Once decoded, though, optical format is one of the most consequential lines on a camera datasheet: it dictates your lens options, your depth of field behavior, and how much of your budget the optics will quietly absorb.
Where the numbers come from
The fraction names the diameter of the vidicon tube that would have produced a similar image area in the 1950s-70s. The actual sensor diagonal is roughly two thirds of the stated figure: a "1 inch" sensor has a 16 mm diagonal (13.2 mm diagonal is the 1" machine vision standard of 13.2 x 8.8... specifically 13.2 mm class), a "2/3 inch" is 11 mm diagonal, "1/1.7 inch" about 9.5 mm, "1/1.8 inch" about 9 mm, "1/2.9 inch" about 6.2 mm. Memorize the order of sizes, not the millimeters: 1" > 2/3" > 1/1.7" > 1/1.8" > 1/2.9".
Format and lenses: the image circle rule
Lenses project circles; sensors sit inside them. The lens image circle must cover the sensor diagonal. Since lens price scales steeply with circle size, sensor format is effectively a proxy for optics cost. The huge installed base of 2/3"-class FA lenses makes 2/3" sensors (like the IMX264 in the MV-CS050-10GM) the economical default. A 1" sensor - the IMX183 in the MV-CS200-10GM - immediately requires 1"-circle optics, a shorter and pricier list. The 1/1.7" IMX226 in the MV-CU120-10GM sits between: compact format, but its 1.85 um pixels demand high-resolution-corrected lenses regardless.
Format and pixels: the resolution trade
Here is the subtle part. Megapixels can be packed onto almost any format; what changes is pixel size. The 2/3" IMX264 spreads 5 MP over larger pixels (3.45 um); the 1/1.7" IMX226 squeezes 12 MP into 1.85 um pixels. Larger pixels gather more photons, tolerate lower-quality optics, and deliver cleaner low-light images. Smaller pixels buy spatial resolution at the price of sensitivity and optical demands. When comparing cameras, always read format and pixel size together - a "more megapixels" claim often just means "smaller pixels crammed onto the same silicon".
Format and geometry: field of view
At the same working distance and focal length, a bigger sensor sees a wider field. That is often the reason to go big: one 1" sensor may replace two tiled 2/3" stations. Conversely, small formats concentrate resolution where narrow fields are fine - code reading, small part verification - and their lenses cost almost nothing.
Format and depth of field
For the same field of view, bigger formats work at longer focal lengths, and longer focal lengths at the same aperture give shallower depth of field. Large-format systems therefore need more careful aperture and lighting discipline to keep tall parts in focus. Small-format systems are forgiving - one reason 1/2.9" and 1/3" sensors dominate simple presence-check readers.
A practical map
1/2.9" to 1/3": code readers, presence checks, fast global shutter work - minimal optics cost. 1/1.8" to 2/3": the machine vision mainstream, 5-6 MP at sensible pixel sizes, vast lens ecosystems - the default choice. 1/1.7": high resolution in a compact format, demanding optics. 1": 20 MP class and above, wide fields, premium lenses, and geometry where a single large image beats tiling.
Every camera on our site lists its optical format and pixel size alongside resolution and frame rate, so the four-way trade - circle, pixel, field, depth - can be evaluated from the product page before any datasheet archaeology.