[in Japanese]

View full abstract

We propose a new spatially variant sampling control system integrated on an image sensor. This sensor can read out spatially variant pixels at high speed, without inputting pixel address for each access. Since it has a memory array that keeps the pixel position to be sampled, the sampling position can be dynamically changed by rewriting the memory array. It can achieve any spatially varying sampling patterns. We show the principles, circuit designs, a prototype of the sensor, and results obtained by the prototype.

View full abstract

A 256 × 256 motion detector on a single chip has been developed. The sensor is fabricated through using the buried charge accumulator and sensing transistor array (BCAST) pixel architecture. The complete on-chip motion detection and the reproduction of high-quality images are performed simultaneously with per-column simple circuits. A 2 μm twin-tub CMOS process is used for its on-chip circuits. A dynamic range of 83.9 dB and a fixed pattern noise (FPN) of 79 dB at 60 frames/ s have been achieved. Furthermore, high-speed operations at 240 frames/s with digital output is possible. These results will extend the application of motion sensing to fields requiring high-quality images.

View full abstract

A new photosensitive field emitter called a “photosensitive floating field emitter” is proposed. It was constructed using an n-type Si field emitter tip and an a-Si : H p-i-n photodiode film. A back incidence type photosensitive floating field emitter was fabricated by depositing intrinsic a-Si : H and p-type a-SiC : H film on the back of a non-gated n-type cone-shaped Si emitter array. The emission current was measured as a function of the illumination intensity and found to increase linearly. The quantum efficiency was about 0.7. However, the emission current tendered to saturate at high illumination level.
To expand the dynamic range of the photodetector, we also examined a photosensitive field emitter composed of a gated field emitter and p-type a-SiC : H/intrinsic a-Si : H photodiode film. Its dynamic range was about 200 times wider than that of non-gated type photosensitive field emitters.

View full abstract

It is well known that the saturation signal level of a CCD image sensor including photo diode with vertical overflow drain (VOD) varies strongly with substrate impurity concentration. However, analytical results between the saturation signal level and substrate impurity concentration have not been obtained yet. In this paper, the relation between the saturation signal level and substrate impurity concentration was calculated using the drift-diffusion model as a function of photo diode structures and operation conditions.By comparing analytical and experimental results obtained from CCD image sensor chips, we found that the impurity fluctuations at the peripherals of CZ, MCZ, and epitaxial wafers were +11%, ±16%, and ±6%, respectively. Those at the center were ±7%, ±9%, and ±6%, respectively. The distinctive features of the impurity fluctuation pattern for these wafers were therefore clarified. The minimum impurity fluctuation level that can be detected is almost ±0.1%.

View full abstract

A new image acquisition system for an HDTV high-speed CCD camera is proposed to meet the increasing demand for HDTV slow-motion shooting equipment. The main idea of this system is to double the frame rate by using two CCDs driven with a half-frame offset. This concept was introduced in a four-chip color separating system. Each 1.3-million-pixel CCD is driven at 1.5 times the pixel-rate of a standard HDTV CCD. Therefore, the green signal acquired with two CCDs has three times the frame-rate of standard HDTV. The red and blue signals also have a tripled frame-rate at an acceptable sacrifice of vertical resolution. This technique has been applied to the development of an HDTV high-speed camera system. The system provided fine HDTV slow-motion images in the hookup for the Nagano winter Olympic games and proved the validity of our image acquisition system.

View full abstract

We have developed a high-resolution color still camera by applying an image-shift of one pixel pitch displacement in the horizontal direction to a color still camera using a Bayer arrangement CCD. This two-positional imaging enables a simplified image-shift mechanism and high-speed imaging. We improved the resolution and suppressed chromatic moire by using a new chromatic moire suppression process that makes use of the correlation between RGB channels. We also eliminated the difference in light quantity between two images using a simple 1-dimensional lowpass filter in the horizontal direction. As a result, we have developed a digital still camera loaded with an image-shift mechanism that can be put to practical use. Our camera has higher resolution than a conventional camera using a Bayer arrangement CCD.

View full abstract

Three structures of the CCM (Correlative Coefficient Multiplying) interpolation method that produce interpolation signals from other color elements were examined for single-imaging-device color cameras. They are (1) vertical interpolation by three of the horizontal element's lines, (2) horizontal interpolation by three of the vertical element's lines, and (3) adaptive interpolation, which selects either the vertical or horizontal interpolation signal, depending on the direction of the object's gradient. The third method perfectly reduced pseudo-colors on a monochromatic object. The SNR of the differential power signal between a 3-CCD camera and single-imaging-device color camera was improved by 8-12 dB compared with the conventional linear interpolation method at monochromatic objects. When an optical low-pass filter was used, the pseudo-color on a monochromatic object was reduced effectively by the first or second method.

View full abstract

The described compression method for images, called VCS, combines traditional color separation and frequency transformation. It produces video images of excellent quality, with reduced chrominance moire and high resolution. An effective application of this method is explained, as is the evaluation of images reproduced by VCS from a color imager with normal complimentary filter.

View full abstract

[in Japanese]

View full abstract