マシンビジョン、自動車用および産業用カメラの画質試験

MTF (Modulation Transfer Function) testing quantifies how well a lens or camera system reproduces contrast at different spatial frequencies. Higher MTF at a given spatial frequency and field position means better contrast reproduction of that detail; comparing sharpness requires the full MTF curve across the field. For machine vision and automotive ADAS applications, MTF directly correlates to the system's ability to resolve small features, read barcodes, or detect objects at distance. We measure lens MTF on our Trioptics HR2 bench and complete camera systems using the slanted-edge (eSFR) method per ISO 12233, providing data at multiple field positions from center to corner, under photopic and 850nm NIR illumination.

MTF testing typically refers to lens-only measurements using specialized optical equipment like our Trioptics HR2, where the lens is tested in isolation without a sensor. eSFR (edge Spatial Frequency Response) testing per ISO 12233:2017 measures the complete camera system including lens, sensor, and image processing. Both produce MTF-equivalent data, but eSFR captures real-world system performance including sensor sampling effects, ISP sharpening, and noise reduction impacts on resolution.

We measure color accuracy using X-Rite ColorChecker charts and calculate Delta E (ΔE) using the CIEDE2000 formula in CIE Lab color space. A ΔE of 1.0 is considered imperceptible to most observers, while ΔE < 3.0 is typically acceptable for consumer applications. For machine vision color sorting or medical imaging, tighter tolerances (ΔE < 1.5) may be required. We test under multiple illuminants (D65, A, TL84) to characterize AWB performance across lighting conditions.

迷光は、レンズ面間の不要な反射、センサーからレンズを介して逆方向に反射する光、および光学コーティングやレンズ筒の表面からの散乱によって生じます。ゴースト像は、特に明るい光源によって二次反射が発生し、センサー上に焦点の合った、あるいは半焦点の像が形成される場合に発生します。自動車用カメラにおいて、これは極めて重要な問題です。なぜなら、1日を通して変化する太陽の角度によってゴーストが発生し、歩行者や道路標識が見えにくくなる可能性があるからです。 当社は、ISO 9358に準拠して、ベールグレア（コントラストの低下）と個別のゴーストの位置の両方を評価しています。

Yes, for surveillance, driver monitoring, and night vision applications, we provide complete image quality characterization at 850nm and 940nm wavelengths. This includes MTF/resolution testing, dynamic range measurement, and sensitivity evaluation under NIR illumination. Many cameras optimized for visible light perform differently at NIR due to lens chromatic focus shift and sensor spectral response variations. NIR testing is essential for systems that must operate in both day and night conditions.

Sample quantities depend on your test plan and statistical requirements. For development testing, 3–5 samples per test condition is typical. For qualification to automotive or military standards, sample sizes are specified by the applicable standard. We can advise on sample planning during the requirements review phase. Keep in mind that image quality testing is generally non-destructive, so samples can be returned after testing for further use.