Specifying the Right Instrument: Balancing Resolution, Speed, and Sensitivity

Q: "Are more pixels always better for an imaging colorimeter?"

"Not necessarily. Higher pixel counts mean larger data volumes, longer readout times, and higher costs—a 61MP device generates over 30x more data per frame than a 2MP unit. Selection should work backwards from application needs: calculate the minimum pixel count based on the DUT's spatial features and FOV, then reserve 20%-50% margin. System-wide speed matching is also critical—without adequate interface bandwidth and processing power, pixel increases won't translate into actual efficiency gains."

Q: "What is the practical impact of the f1' value on measurement results?"

"The f1' value quantifies the deviation between the filter's spectral response and the CIE theoretical curves; lower values mean better matching. For broadband sources like white LEDs, f1' of 3%-5% is usually acceptable. However, for narrowband sources like monochromatic LEDs or OLED subpixels, an instrument with f1'=5% measuring a narrowband red LED may exhibit color coordinate deviations of Δx,y≈0.01—unacceptable in high-end display inspection. Filters with f1'≤2% or spectral correction algorithms are then required."

Q: "Why do HUD and AR/VR testing require specialized lenses?"

"HUD and AR/VR devices generate virtual images rather than physical display surfaces, so the measurement system must simulate human eye optical characteristics. The key configuration is a conoscopic lens, which captures light from various angles at a fixed position, simulating how a human eye observes a HUD virtual image within the eyebox. Standard lenses cannot perform this measurement. AR/VR devices with FOV exceeding 100 degrees also require ultra-wide-angle or fisheye front-end optics."

Table of Contents

Radiant Vision Systems ProMetric I series imaging colorimeter product array—Imaging colorimeters of different pixel grades to meet diverse application needs (Image Source: Radiant Vision Systems) — 图片来源于网络，如有侵权，请联系删除

Introduction: Selection as an Optimal Solution under Constraints
#

An Imaging Colorimeter is not a “one-size-fits-all” tool. Different measurement scenarios pose completely different requirements for the core parameters of the instrument: smartphone screen inspection requires high spatial resolution to resolve individual sub-pixels; large-size TV panel inspection requires a wide field of view (FOV) to cover the entire panel; MicroLED inspection requires extreme dynamic range to handle the luminance differences of self-luminescent chips; and HUD or AR/VR testing requires special optical configurations to simulate human viewing conditions.

The core task of selection is to find the optimal balance between resolution, speed, sensitivity, dynamic range, and cost for your specific application. Over-specification leads to unnecessary cost waste, while under-specification results in measurement data that fails to meet quality standards—both are failures in engineering decision-making.

This article systematically combs through the core parameter dimensions of imaging colorimeter selection and provides configuration recommendations for typical application scenarios, offering an actionable methodology for engineers and procurement decision-makers.

I. Analysis of Core Parameter Dimensions
#

Lens selection guide for ProMetric I series imaging colorimeters—Different focal lengths for various FOV and resolution needs (Image Source: Radiant Vision Systems / DirectIndustry) — 图片来源于网络，如有侵权，请联系删除

1.1 Spatial Resolution: More Pixels are Not Always Better
#

The spatial resolution of an imaging colorimeter is determined by the effective pixel count of its sensor, with common market configurations ranging from approximately 2 megapixels (2MP) to 151 megapixels (151MP). Pixel count directly determines the smallest spatial feature size distinguishable at a given FOV.

Key Concept: Sensor pixels per measured pixel (DUT pixel). This ratio determines spatial measurement accuracy. In display panel Demura applications, it is typically required that at least 3-5 sensor pixels cover one display pixel to accurately extract its luminance and chromaticity. Higher sensor pixel densities (e.g., 8-10 sensor pixels per display pixel) provide more reliable sub-pixel level measurements.

However, higher pixel counts also mean larger data volumes, longer readout times, and higher costs. A 61MP imaging colorimeter generates over 30 times more data per frame than a 2MP device, placing significantly higher demands on data transmission bandwidth, storage capacity, and processing speed.

1.2 Dynamic Range: Key to Measurable Luminance Span
#

Dynamic Range defines the ratio between the highest and lowest luminance an instrument can distinguish in single or multiple exposures. Single-exposure dynamic range is usually determined by sensor bit depth and dark current noise, typically ranging from 59-76 dB. Through multi-exposure High Dynamic Range (HDR) synthesis, system dynamic range can be extended to 120-140 dB.

Dynamic range is critical in:

OLED/MicroLED Black State Measurement: Self-luminescent displays may have residual emission even when pixels are “off,” requiring measurement at extremely low luminance.
Contrast Measurement: Scenarios where high-brightness and dark areas coexist in the same frame.
Automotive Ambient Lighting: Weak emission in dark environments requires high sensitivity, yet high-brightness modes must not saturate.

1.3 Spectral Matching Accuracy: Engineering Significance of f1’ Values
#

Imaging colorimeters simulate the CIE standard observer’s tristimulus functions x(λ), y(λ), and z(λ) using filters. The deviation between the actual spectral transmittance of the filters and the theoretical curves is quantified by the f1’ value (Spectral Mismatch Index); a smaller f1’ indicates more accurate matching.

Typical industrial-grade imaging colorimeters have f1’ values for V(λ) between 1.5% and 5%. For broadband light sources (e.g., white LEDs), f1’ in the 3%-5% range is usually acceptable. However, for narrowband sources (e.g., monochromatic LEDs, RGB sub-pixels of OLEDs), a large f1’ leads to significant chromaticity measurement errors.

Practical Impact: An instrument with f1’=5% measuring a narrowband red LED might exhibit a color coordinate deviation of Δx,y ≈ 0.01, which is unacceptable in high-end display inspection. In such cases, filters with lower f1’ (e.g., f1’ ≤ 2%) or spectral correction algorithms are required.

1.4 Measurement Speed: The Hard Constraint of Takt Time
#

In production line environments, measurement speed directly impacts Takt Time. Speed is determined by:

Exposure Time: Low-luminance measurements require longer exposures; for example, measuring black states at the 0.1 cd/m² level may take several seconds.
Multi-frame Synthesis: HDR mode requires capturing and synthesizing multiple images at different exposures, each extra frame extending the total time.
Data Readout and Transmission: High-pixel sensors take longer to read out; for instance, readout for 61MP is significantly longer than for 2MP.
Image Processing: Applications like Demura require extensive calculations after capture.

Typical production line requirements demand a complete measurement (exposure, readout, transmission, and basic processing) within 3-15 seconds. Selection must confirm that the configuration can complete all required tasks within the target Takt Time.

1.5 Sensor Cooling: Dark Current Suppression and Lower Detection Limit
#

Sensors generate Dark Current during operation—thermally generated charges accumulate in pixels even without light. Dark current magnitudes are exponentially related to sensor temperature: every 6-8°C decrease approximately halves the dark current.

For applications requiring measurement of low-luminance targets (e.g., OLED black states, display light leakage), sensor cooling is indispensable. Typical solutions include:

Passive Cooling (No Active Cooling): Suitable for high-luminance scenarios; simple and low-cost.
Single-stage Thermoelectric Cooling (TEC): Lowers sensor temperature by ~15-25°C below ambient; suitable for most production line applications.
Multi-stage Thermoelectric Cooling: Lowers temperature to -20°C to -40°C or lower; suitable for scientific-grade extremely low-luminance measurement.

Cooling adds size, weight, power consumption, and cost, and requires a startup time (typically 1-5 minutes) to stabilize. Decision on cooling should be based on the required lower luminance limit.

II. Application Scenarios and Configuration Decision Matrix
#

Product appearance of ProMetric I series imaging colorimeter—Industrial-grade design for both production line and laboratory needs (Image Source: Radiant Vision Systems / Photonics) — 图片来源于网络，如有侵权，请联系删除

2.1 Smartphone Screen Inspection
#

Core Needs: High resolution, moderate speed.

Typical smartphone resolutions range from 1080x2400 (FHD+) to 1440x3200 (QHD+), with millions of sub-pixels. Achieving pixel-level Demura and defect detection requires high imaging resolution.