A High-Efficient Measurement System With Optimization Feature for Prototype CMOS Image Sensors.

In this paper, a gray-scale CMOS image sensor (CIS) characterization system with an optimization feature has been proposed. By using a very fast and precise control of light intensity, based on the pulsewidth-modulation method, it is avoided to measure the illuminance every time. These features accelerate the multicriteria CIS optimization requiring many thousands of measurements. The system throughput is 2.5 Gb/s, which allows for capturing images from large arrays of the size $3000 \times 3000$ pixels at the rate of 25 frames/s, or small arrays ($128 \times 128$) at the rate of 15 000 frames/s. The efficient transfer of measurement data to the external software allows immediate presentation of optimization results in 3-D plots. The system automatically measures nonuniformity, spatial noise, temporal noise, signal-to-noise ratio, dynamic range, nonlinearity and image lag. A flat diffuser has been proposed as a cheaper alternative to an integrating sphere. This optical front-end is particularly useful for testing prototype CISs and vision-chips implemented in standard CMOS technologies as low- or medium-dynamic-range imagers. [ABSTRACT FROM AUTHOR]