SMEAR COMPENSATION FOR MOVING OBJECTS
Publication Date: 2005-Apr-08
The IP.com Prior Art Database
A method for reducing unwanted smear signal within an image sensor includes the steps of: providing the image sensor with a matrix of pixels arranged in a plurality of rows and columns with a vertical shrift register allocated for each of the columns and at least one horizontal shift register operatively coupled to the vertical shift registers, wherein each of the columns of pixels are formed with the vertical shift registers having a plurality of phases allocated for each of the pixels and a plurality of gate electrodes of the vertical shift register for each of the pixels, and clocking means for causing the transfer of charge form the pixels to the vertical shift registers and through the horizontal registers; applying, at a first time period, a first set of voltages to the phases of the gate electrodes of the vertical shift registers sufficient to transfer charge from the photodiode to the vertical register; applying during a second time period a sequence of voltage pulses to the vertical registers that transfers the photoelectron signal to a storage area; during this second time period, rows of overclock signal at the bottom and top of the image are also clocked into the storage area and these overclocked rows contain unwanted smear signal from both stationary and moving bright objects; during a third time period the overclock signal and image are clocked row by row into a horizontal register which is read out as video data; the unwanted smear signal in the overclocked rows at the top and bottom of the image are used to correct the undersirable smear streaks in the raw image data.
Smear Compensation for Moving Objects
Field of the Announcement:
The present disclosure relates to charge coupled devices (CCDs) imager systems incorporating improved smear reduction techniques.
Background of the dislosure:
Under normal operation a CCD transfers an image row by row into a horizontal register where each row is sequentially read out. If one or more bright objects are in the image then spurious signal can leak into the registers while the image transfer takes place resulting in unwanted streaks. Figure 1a shows an image with a bright incandescent lamp. The bright filament results in a smear streak labeled 1 in Fig 1a. If the lamp were still during image readout then the smear streak would run from top to bottom. However, the lamp moved slightly during readout and therefore there is a slight tilt to the streak. In other words the streak is no longer perpendicular to the horizontal register. This complicates any smear subtraction algorithm. A similar problem occurs if the brightness of the object changes during readout. This results in a gradation from top to bottom in the streak. The obvious solution to this problem is to utilize a shutter to prevent a projected image form reaching the CCD imager during the transfer portion of each successive field. However, due to cost and size constraints this is not a preferable approach for many cameras. This announcement describes how to remove the unwanted smear streaks by using a specific timing sequence and subtraction algorithms.
Much of what is described here can be found in U.S. Pat. No. 4,490,744 entitled “Smear Reduction Technique For a Field-Transfer Imager System” by Peter Levine. This patent applies to full frame / frame transfer CCD imagers while this announcement applies to interline CCD imagers and frame interline transfer CCD imagers.
Detailed description of the disclosure:
Figure 2 shows the layout for a specific frame interline transfer CCD with 3040 x 1620 active pixels. Operation of this sensor is as follows: 1) photosignal is collected as electrons in each pixel for fixed integration time; 2) the charge is transferred into the vertical charge coupled device (VCCD) registers; 3) the signal is rapidly clocked to the storage area (it is during this portion of the readout that the smear streak occurs); 4) signal in the storage area is then slowly clocked row by row into the horizontal register where it is readout by an output amplifier at the end of the horizontal register.
A row by row timing sequence for this sensor is illustrated in Figure 3. The advantage of this timing sequence is that for each column there are 8 overclock pixels at the start of each column and at the end of each column. Of the 8 overclock pixels, 4 provide dark current signal (labeled D) and 4 provide dark current plus smear signal (labeled S). Using the information provided by these pixels the background dark current signal can be subtracted off for each column. Also, the s...