Most machine vision lines live happily at 2-5 megapixels. But some jobs keep pushing you toward the high end: reading laser-marked 2D codes on tiny components, inspecting solder joints on a dense PCB, or flagging scratches on a polished metal surface the eye can barely see. Once you go shopping above 10 megapixels, Hikrobot's GigE lineup gives you two very different answers: the 12-megapixel MV-CU120-10GM family and the 20-megapixel MV-CS200-10GM. More megapixels sounds like an automatic win. It usually isn't. Here is how the two actually compare.
The two candidates
Both are monochrome GigE area scan cameras built around Sony CMOS sensors, and both are C-mount. That is where the similarity ends.
| MV-CU120-10GM | MV-CS200-10GM | |
|---|---|---|
| Sensor | Sony IMX226 | Sony IMX183 |
| Optical format | 1/1.7" | 1" (13.2 mm diagonal) |
| Resolution | 4024 x 3036 (12.2 MP) | 5472 x 3648 (19.9 MP) |
| Max frame rate | 9.7 fps | 5.9 fps |
| Interface | GigE | GigE |
If you want the same 20 MP class in color, the MV-CS200-10GC uses the identical IMX183 sensor with a Bayer filter. And if the 12 MP route appeals but the budget is tight, the MV-CE120-10GM shares the IMX226 sensor and nearly the same frame rate (9.6 fps) in a lighter, more affordable housing.
Do the sampling math first
The honest way to pick a resolution is to stop counting pixels and start measuring microns. Divide your field of view by the horizontal pixel count; that number, not the megapixel figure, decides what you can actually detect.
Take a 200 mm wide field of view. The 12 MP camera gives you 200 mm / 4024 = about 49.7 um per pixel. The 20 MP camera gives 200 mm / 5472 = about 36.5 um per pixel. For reliable defect detection you generally want the smallest defect you care about to span at least 3 pixels, which puts your practical detection floor around 150 um with the 12 MP camera and 110 um with the 20 MP camera - on that same field of view.
Notice what happened there: 64% more pixels bought you only about 26% finer detail. That is the megapixel trap. If your defect spec is 200 um or larger, both cameras pass and the 12 MP option will run faster and cost less. If you truly need sub-120 um detection at that FOV, the 20 MP earns its price.
Frame rate: the quiet trade-off
Higher resolution on the same GigE link means slower frames. The 12 MP camera manages 9.7 fps against 5.9 fps for the 20 MP - and because your cycle budget is per part, the gap compounds with motion.
Concretely: on an indexing table presenting one part per second, both cameras are comfortable. On a conveyor at 60 parts per minute, the 20 MP camera is already the bottleneck (5.9 fps means you cannot inspect more than ~354 parts per minute even in theory, and real trigger jitter eats into that). Neither camera belongs on a high-speed bottling or blister line - that is continuous-scan or line-scan territory, not area-scan high resolution.
One practical note: both cameras are mono (GM suffix). For measurement tasks, code verification, and most surface-defect work, mono is the right call - it carries roughly three times the signal-to-noise of the equivalent Bayer color sensor and skips demosaicing artifacts entirely. Choose the color GC variants only when hue itself carries information: printed labels, wire color verification, food sorting.
Optics: the cost nobody budgets
The IMX183 in the 20 MP camera is a 1-inch sensor; the IMX226 is 1/1.7". That difference reaches straight into your lens invoice. A 1" format demands a lens with a matching image circle, corrected across 5472 tiny 2.4 um pixels - that is a serious optical design, and cheap C-mount FA lenses will resolve the center of the frame while the corners quietly smear.
The 12 MP camera's smaller sensor and 1.85 um pixels are friendlier: ordinary 5-megapixel-class FA lenses often hold up acceptably at f/4-f/5.6, though a lens matched to the format always measures better. Whichever route you take, mount the actual lens and shoot an actual target before committing - MTF charts do not argue with a USAF resolution chart on your own part.
A quick decision guide
| Your situation | Pick |
|---|---|
| Defect spec >= 150 um at moderate FOV, faster cycle | 12 MP - MV-CU120-10GM |
| Same as above but cost-sensitive | 12 MP - MV-CE120-10GM |
| Sub-120 um detail needed, slower line | 20 MP - MV-CS200-10GM |
| Color matters: print, labels, sorting | 20 MP - MV-CS200-10GC |
| Tiny defects but huge FOV (200 mm+) | Consider stepping up to 25 MP-class or line scan instead |
New or refurbished?
One more practical option: both sensor generations have been on the market long enough that refurbished stock is widely available. A refurbished MV-CU120-10GM typically lands around 30% below the new unit and carries a shorter but real warranty - sensible for a pilot line or a university lab. For a 24/7 production line where a camera swap means stopping the belt, buying new with a 12-month warranty is usually the less expensive decision over three years.
We stock both resolution classes with new and tested refurbished units, ship worldwide free of charge, and the specification tables on each product page list the exact sensor, frame rate, and interface data pulled from the manufacturer datasheets - no marketing rounding.