CMOS prototype vision chip with digital pixel structure for grey level image segmentation by means of thresholding and time multiplexing

In this article, a CMOS prototype vision chip with digital pixel structure for grey level image segmentation by means of thresholding and time multiplexing is presented. This approach splits scenes into m frames (one frame per grey level interval). One advantage about this design is that an analogue to digital converter is not required. Moreover, image acquisition and segmentation are performed at the same time by pixels that work simultaneously with each other. The performance from each pixel deals with a maximum quantum efficiency of 0.65, pixel size of 132 μm × 176 μm, fill factor of 0.78%, dark current of 15mV/s, power dissipation per frame of 341W, minimum exposure time of 28.6s, maximum exposure time of 1.9ms, random noise of 3mV, optical dynamic range of 51dB and majority of cells with 0-3% of mismatch. Scene decomposition into 256 images occurs in 30.5ms with white illumination of 650 Lx.