Abstract: A process for performing white balancing of an image is performed by subdividing an image into a plurality of subframes, and then analyzing each subframe to determine if that subframe is predominantly monochromatic other than gray. If so, that subframe is excluded from the computation of the gain adjustments in the white balancing operation. As a result, the white balance process is performed using only the multicolored and/or gray subframes, thus allowing the overall white-balance of the image to be shifted only when a change in the color average is due to a change in the spectra of illumination, and not a presence of large monochromatic areas in the image.

Abstract: The invention provides a new method and apparatus for NTSC and PAL image sensors which employs fusion of adjacent row pixel charge samples to generate image data for a row. A variety of fusion schemes are possible for fusing the pixel signals from the adjacent rows. The rows of pixels are scanned so that each scan takes an odd row signal sample and, in some cases, an adjacent even row signal sample when specified conditions are met. One sampled row of the two adjacent rows integrate an image with a first integration period while the other adjacent row integrates an image with a second integration period.

Abstract: An image processing method and system for a multi-sensor network camera. The method and system including generating a plurality of full resolution images in Bayer array format (Bayer images) produced by a plurality of image sensors; interpolating a plurality of low resolution Bayer images from the full resolution Bayer images during the readout of the full resolution images from the sensors, storing the full resolution Bayer images and the interpolated low resolution images in a plurality of buffer memories, respectively and without demosaicing the full resolution Bayer images, during the readout of the full resolution Bayer images from the image sensors, by respective plurality of pre-processors; demosaicing the plurality of low resolution Bayer images to generate a corresponding plurality of low resolution demosaiced images, by an image post processor; and transmitting the plurality of low resolution demosaiced images over a computer network to a user for viewing.

Abstract: An image processing system and method compares each pixel of an image obtained from an image sensor array with at least eight surrounding pixels of the same color in the filter array. If the signal of a given pixel is larger than the respective signals of all eight surrounding pixels of the same color, then the value of that central pixel signal is substituted with the maximum signal value among the surrounding eight pixels of the same color. Similarly, if the signal of a given pixel is smaller than the respective signals of all eight surrounding pixels of the same color, then the value of that central pixel signal is substituted with the minimum signal value among the surrounding eight pixels of the same color.

Abstract: An apparatus and method for measuring the breakpoint of a response curve representing the voltage output of an image array having an extended dynamic range. By flooding a light-opaque pixel with a charge and then applying an intermediate reset voltage to the pixel, the signal is read from the pixel and stored. The full reset voltage is applied to the pixel, and then the signal in the pixel is read and stored. The voltage output difference is the difference between the first and second stored signal. The voltage output difference is then used to determine the voltage of the knee point. Further, a conventional saturated pixel can be reset with an intermediate reset just prior to readout. The resulting signal can then be used to determine the voltage of the knee point.

Abstract: In an image processing system, the imager gain and the exposure time are adjusted based on a predefined stepping sequence using a stepping table designed to maximize the signal-to-noise ratio in the image. This is achieved by providing a stepping sequence with each step having the largest suitable integration time (exposure) and an appropriate amplifier and digital gain setting, while achieving an equal relative percentage change in image brightness between adjacent sequence steps. The size of the executed AE steps is proportional to the distance between a current image luminance and the target luminance. This is achieved by the capability to skip one or more steps in the stepping sequence, if appropriate, for each relevant executed step, with the number of skipped steps, if any, being proportional to the magnitude of the deviation from the target brightness.

Abstract: A multi-sensor network camera providing up to 360 degrees angle of view. The includes multiple image sensors with individual optics, one or more image processors, compression units and network interfaces mounted in the single housing. The image sensors are positioned in non-parallel planes, cumulatively providing panoramic field of view and image streams originating from all sensors share the same image compression and network interface hardware, providing for low cost implementation. The images from all sensors are transmitted over the network simultaneously via packet interleaving, with appropriate bandwidth reduction achieved by image decimation. Simultaneously with transmission of decimated images from all sensors, full resolution window or entire image of one or more sensors may also transmitted, where the selection of contents is based on motion detection or user setting.

Abstract: The subject of this invention is the network camera comprising an image sensor, image processor, and network interface, where image processor is capable of detecting the motion in the field of view of the camera, extracting/windowing the portion (sub-window) of the image that corresponds to that motion and submitting thus identified sub-window to the network interface hardware for transmission and where said image processor is also capable of submitting to the network interface the full field of view reduced-resolution (decimated) image either as the only image to be transmitted or as the image to be transmitted in a time-interleaved fashion with the said image window containing the motion.

Abstract: A method for determining a correction for a spectra of scene illumination based on analysis of color content of a video image, in which average image colors are obtained by considering only those image pixels that have significantly different color from any of its neighboring pixels resulting in selection of only those pixels that lay on the boundaries of monochrome image segments, while avoiding their inner areas. Therefore, the white balance method is not only tolerant to the presence of large non-gray monochrome areas in the image, but also avoids the image areas that have small chrominance values and appear to be gray due to specific illumination spectra, while not actually corresponding to the grey areas of the scene.

Abstract: A system of correcting for color filter array by correcting for a plurality of colors in the color filter array. A color correction matrix is formed which corrects each color for a plurality of desired characteristics simultaneously. These desired characteristics can include, for example, all 24 colors in a chromaticity chart. An additional aspect includes weighting the more important colors relative to the less important colors.

Abstract: The subject of the present invention is the network video camera capable of delivering high-resolution digital images, comprising over a million of pixels per image, at full motion video frame rate.

Abstract: A video system including one or more high resolution network video cameras, a computer and a monitor. The video cameras are capable of simultaneously generating a full field of view reduced-resolution of an image and a full-resolution sub-window of the image. The computer is configured to receive images from the video cameras via a computer network and display the images on the monitor in which a full field view of the image and a sub-window of the image are displayed on the monitor simultaneously.

Abstract: A CMOS imager in which a CMOS image sensor, a color image processing module and an image compression module are all provided on a single die. Both the color image processing module and the image compression module incorporate pipelined architectures to process the image data at a video rate in a massively parallel fashion.

Abstract: The integration time for an imager is adjusted to be an integer multiple of the light intensity received by the imager to reduce or eliminate flickering.

Abstract: A network camera comprising an image sensor, an image processing unit, a buffer memory and a network interface. The image processing unit is configured to divide the entire image readout by the image sensor into a plurality of overexposed regions and a plurality of non-overexposed regions, select a subset of the overexposed regions containing most oversaturated pixels, control said image sensor to readout only said selected subset of the overexposed regions with adjusted optical integration time, apply multiplicative scaling to readout pixels of said selected subset of the overexposed regions to generate scaled pixels, and replace pixels in the entire image corresponding to said selected subset of the overexposed regions with respective said scaled pixels to increase dynamic range of the entire image.

Abstract: The subject of this invention is the network camera sensitive to both visual and near-infrared or infrared spectral bands and comprising two or more image sensors with different spectral responses, one or more image processors, and network interfaces. In the preferred embodiment the network camera disclosed in this invention comprises two image sensors with individual optics, where both sensors have similar fields of view. In the preferred embodiment of the present invention one of the sensors is color image sensor with infrared-cut filter positioned in front of pixel array, while the other sensor is monochrome image sensor with optics characterized by lower f-number than that used for color sensor and without infrared-cut filter in its field of view. In one of the embodiments of this invention, monochrome sensor has larger pixel sizes than color sensor to provide for higher low-light sensitivity.

Abstract: An apparatus and method for measuring the breakpoint of a response curve representing the voltage output of an image array having an extended dynamic range. By flooding a light-opaque pixel with a charge and then applying an intermediate reset voltage to the pixel, the signal is read from the pixel and stored. The full reset voltage is applied to the pixel, and then the signal in the pixel is read and stored. The voltage output difference is the difference between the first and second stored signal. The voltage output difference is then used to determine the voltage of the knee point. Further, a conventional saturated pixel can be reset with an intermediate reset just prior to readout. The resulting signal can then be used to determine the voltage of the knee point.

Abstract: An image processing method and system for a multi-sensor network camera. The method and system including generating a plurality of full resolution images in Bayer array format (Bayer images) produced by a plurality of image sensors; interpolating a plurality of low resolution Bayer images from the full resolution Bayer images during the readout of the full resolution images from the sensors, storing the full resolution Bayer images and the interpolated low resolution images in a plurality of buffer memories, respectively and without demosaicing the full resolution Bayer images, during the readout of the full resolution Bayer images from the image sensors, by respective plurality of pre-processors; demosaicing the plurality of low resolution Bayer images to generate a corresponding plurality of low resolution demosaiced images, by an image post processor; and transmitting the plurality of low resolution demosaiced images over a computer network to a user for viewing.

Abstract: The invention provides a new method and apparatus for NTSC and PAL image sensors which employs fusion of adjacent row pixel charge samples to generate image data for a row. A variety of fusion schemes are possible for fusing the pixel signals from the adjacent rows. The rows of pixels are scanned so that each scan takes an odd row signal sample and, in some cases, an adjacent even row signal sample when specified conditions are met. One sampled row of the two adjacent rows integrate an image with a first integration period while the other adjacent row integrates an image with a second integration period.

Abstract: The invention provides a new method and apparatus for NTSC and PAL image sensors which employs fusion of adjacent row pixel charge samples to generate image data for a row. A variety of fusion schemes are possible for fusing the pixel signals from the adjacent rows. The rows of pixels are scanned so that each scan takes an odd row signal sample and, in some cases, an adjacent even row signal sample when specified conditions are met. One sampled row of the two adjacent rows integrate an image with a first integration period while the other adjacent row integrates an image with a second integration period.