JP4437940B2 - Television camera - Google Patents

Description

  The present invention relates to a television camera, and more particularly to improvements in CCD accumulation interpolation processing and cyclic noise reduction processing.

  Conventionally, when an electronic zoom process for enlarging an image with a television camera is realized, the output image of the CCD image sensor is subjected to analog processing such as sample & hold processing (CDS) and automatic gain adjustment (AGC), and analog signal processing. After digital conversion (A / D) processing and digital processing (DSP) such as gamma correction and contour correction, video signals of all imaging areas are once recorded in the field memory, and desired after 2 fields (1 frame) A portion of the video data is cut out, and the video is electrically enlarged by performing pixel enlargement processing and interpolation processing between adjacent pixels.

  As for the accumulation interpolation processing, the signal charge multiplied by increasing the exposure time of the CCD image sensor is taken out, the video signal is recorded in the field memory, and the field for the video missing field period during exposure is recorded. By reading out video signal data several fields before recorded in the memory, continuous high-sensitivity video signals are obtained.

In addition, with regard to the cyclic noise reduction processing, by adding the video signal after A / D conversion and the video signal of one field before at a certain ratio, the video signal component having a high correlation between fields is not greatly affected. The noise is reduced by using the fact that the noise component whose level constantly fluctuates for each field is averaged and becomes smaller.
Here, when these three functions are to be realized at the same time, the accumulation interpolation processing and the cyclic noise reduction processing are generally performed before the digital signal processing (DSP), and the electronic zoom processing is generally performed after the DSP. It is. The reason is that if a large amount of noise components are included when performing gamma correction or contour correction in digital signal processing, the noise components will be excessively emphasized, so that it is necessary to perform noise reduction processing before the DSP, This is because if the electronic zoom processing is performed before the accumulation interpolation processing, the zoom magnification cannot be changed during the field missing period, and the tracking of the magnification change is delayed.

An example in which an electronic zoom function and a storage sensitivity increasing function are used together in a single-panel television camera is disclosed in Patent Document 1.
JP 2000-287124 A

  In the above-described conventional technology, the circuit for performing the accumulation interpolation process and the cyclic noise reduction process and the circuit for performing the electronic zoom process are located apart from each other with the DSP interposed therebetween, so that a field memory and its control circuit are required, respectively. The circuit scale and parts cost will increase. Also, when video level detection for automatic gain adjustment or lens iris control is performed by the DSP, a difference occurs between the detection area of the DSP and the area enlarged by the subsequent electronic zoom, so the center of the shooting area is dark, for example. When there is a bright subject in the corner, the DSP adjusts the level of the bright subject and the automatic gain adjustment and lens iris control do not operate even if the camera angle of view changes with electronic zoom. The problem that the brightness of the camera does not change occurs.

  An object of the present invention is to realize a space-saving and low-cost television camera that can independently control video data interpolation, noise reduction, and video enlargement functions when the storage sensitivity is increased, in order to solve the above-described problems. is there.

  In order to achieve the above object, the television camera of the present invention includes an electronic zoom process for electrically extracting and enlarging a part of a photographed video signal, and an intermittent video signal accumulated for a long time by a CCD image sensor. Are interpolated in the memory to obtain a continuous image, and a cyclic noise reduction process for adding the output image of the CCD image sensor and the image one field before is performed.

  In addition, the present invention provides a television camera that performs accumulation sensitivity increase, noise reduction, and image enlargement, and includes a memory for storing video data and a computing unit, and interpolates video data when the storage sensitivity increases, reduces noise, and The image enlargement is performed by the memory read control and the arithmetic unit.

  Furthermore, the present invention multiplies the imaging device that outputs video signals by exposing incident light to one or more fields, a digital converter that converts the video signals into digital data, and the digital data and the first or second data. A first multiplier, an adder for adding the output data of the first multiplier and the third data, a memory for recording the output data of the adder, and 2 after being recorded from the memory At least a second multiplier that multiplies the field-delayed data with the fourth or fifth data and outputs the third data, and records the data delayed from the memory by one field after being recorded from the memory It is a signal.

  According to the present invention, it is necessary to realize accumulation interpolation processing, cyclic noise reduction processing, and electronic zoom processing using a set of memory, a memory control circuit, a multiplier, and an adder arranged in front of the DSP. By sharing most of the circuits, a space-saving and low-cost television camera can be realized. Furthermore, the above three processes can be controlled independently, so that all eight combinations can be used. .

The operation of one embodiment of the television camera according to the present invention will be described below with reference to FIG. FIG. 1 is a block diagram of a television camera showing an embodiment of the present invention.
In FIG. 1, 1 is a color separation prism that decomposes incident light (not shown) into three primary color lights of red (R), green (G), and blue (B), and 2a to 2c are electric charges obtained by photoelectric conversion. Image pickup devices (CCD) 3a to 3c for sequentially transferring the video signals and outputting the video signals. The automatic gain control circuits (CDS & AGC) 4a for sampling and holding the video signal outputs of the CCDs 2a to 2c and amplifying them to a predetermined level -4c are analog-to-digital converters (A / D) for converting analog video signals to digital signals, 5a-5f are multipliers, 6a-6c are adders, 7 is a field memory, 8 is a memory control circuit, and inputs The output port 8b is shared by the port 8a, the accumulation interpolation circuit, and the cyclic noise reduction circuit, and the output port 8c is dedicated for electronic zoom. Reference numeral 9 denotes a zoom interpolation circuit that smoothly interpolates between pixels expanded by electronic zoom by mixing adjacent pixel data, and 10 denotes various types for improving image quality such as gamma correction, contour correction, and tone correction. DSPs 11a to 11b that perform digital signal processing of the digital-analog converter (D / A) that converts the luminance signal (Y) and color signal (C) encoded from the R, G, and B signals by the DSP 10 into analog signals. ), 12 is an adder that multiplexes the luminance signal and the color signal into a color video signal, 13 is a connection terminal for connecting the color video signal to a monitor (not shown), and 14 is an accumulation and readout of the CCDs 2a to 2c. A CCD drive signal generator 15 for generating a timing signal for the camera, and 15 is a reference signal of the CCD drive signal generator 14 by counting the vertical synchronization signal of the television camera. The field counter 16 switches the coefficients of the multipliers 5a to 5f in synchronization with the reference signal, and 16 sets data in the field counter 15 and the memory control circuit 8 so as to have a predetermined accumulation time and electronic zoom magnification. It is a microcomputer (CPU).

  Next, the accumulation interpolation process when accumulating for four field periods will be described. The video signal, which is sensitized four times as usual, is read out from the CCDs 2a to 2c once every four fields by the drive pulse of the CCD drive signal generator 14, and is multiplied through the CDS & AGCs 3a to 3c and A / D 4a to 4c. Input to the devices 5a to 5c. Here, in synchronization with the CCD drive signal, the coefficients of the multipliers 5a to 5c connected to the output side of the A / D 4a to 4c and the multipliers 5d to 5f connected to the memory output port 8b change, and the adders 6a to 6c in the next stage change. The selector is composed of a combination with. The reason for not using the selector is to share it with the cyclic noise reduction circuit described below. FIG. 2 shows a timing chart from the A / D 4a to 4c output to the input of the zoom interpolation circuit 9 that performs pixel interpolation. The horizontal axis represents the elapsed time in units of vertical synchronization signals (fields). In accordance with the timing at which video is output to the A / Ds 4a to 4c, the coefficients of the multipliers 5a to 5c are set to 1 times, the coefficients of the multipliers 5d to 5f are set to 0 times, and the accumulation period in which the video is missing Are set to 0 and 1 respectively. Accordingly, the output data of the A / D 4a to 4c is output during the video output period of the CCD and the output data of the memory 8b is output during the accumulation period of the CCD. Here, in the memory output port 8b, the data written to the field memory 7 from the memory input port 8a is read one field later, so that the output of the adders 6a to 6c includes the data of the video output period of the CCDs 2a to 2c. Thus, the video missing period being accumulated is interpolated, and the video of the same field is continuous for four fields. In the other memory output port 8c, since the data in the memory 7 is read out after two fields, the interpolated data is delayed by two fields and output. The reason for reading after two fields is that the CCD is interlaced, so the scanning fields of the CCDs 2a to 2c and the reading field of the field memory 7 must be matched. In the arrows in the figure, the solid line represents data movement through the field memory 7, and the broken line represents data movement through the multipliers 5a to 5f and the adders 6a to 6c.

  Next, a case where a cyclic noise reduction process is used in combination with this accumulation interpolation process will be described. FIG. 3 is a timing chart when the number of accumulated fields is 4 and the cyclic coefficient is 0.5. The coefficients of the multipliers 5a to 5c are set to 0.5 times and the coefficients of the multipliers 5d to 5f are set to 0.5 times in accordance with the timing when the images are output to the A / Ds 4a to 4c. The accumulation periods are set to 0 times and 1 time, respectively. Therefore, at the outputs of the adders 6a to 6c, the video interpolated with the addition average data of the outputs of the A / Ds 4a to 4c and the output data of the field memory 7 is continuous for four fields. Since addition averaging is performed between fields, only randomly changing noise can be reduced without affecting a video signal having a high correlation between fields. As the cyclic coefficient has a higher ratio on the multipliers 5d to 5f side, data is added over more fields, so that a higher noise reduction effect is obtained.

  Next, the case where the electronic zoom process is used in addition to the accumulation interpolation process and the cyclic noise reduction process will be described with reference to the timing chart of FIG. The signal path to the memory input port 8a is the same as described above, but zoom data obtained by partially cutting out and reading out the output data of the adders 6a to 6c recorded in the memory is output to the output of the memory output port 8c. Is done. The signal waveform of the memory output port 8c in FIG. 4 is obtained by partially cutting out the central portion of the output waveform of the adders 6a to 6c and enlarging the video. In the memory input port 8a, horizontal scanning of the imaging area is associated with a row in the address space of the field memory, and vertical scanning is associated with a column in the address space of the memory and recorded in order in the memory. If the column address is read twice in succession, the zoom is doubled. Further, since all the shooting areas are recorded in the memory, it is possible to zoom in not only on the center of the shooting area but also on any area by changing the read address range. Needless to say, when the zoom magnification is set to 1, it is the same as the combination of the accumulation interpolation process and the cyclic noise reduction process.

  Although the present invention has been described in detail above, the present invention is not limited to the three-plate system television camera described herein, and imaging other than the above-described single-plate system, multiple-plate system, imaging tube system, etc. Needless to say, it can be widely applied to the method.

The block diagram of the television camera which shows one Example of this invention. The timing chart for demonstrating the operation | movement of the accumulation | storage interpolation process of one Example of this invention. The timing chart for demonstrating operation | movement at the time of using together the accumulation | storage interpolation process and cyclic | annular noise reduction process of one Example of this invention. The timing chart for demonstrating the operation | movement at the time of using together the accumulation interpolation process of one Example of this invention, a cyclic | annular noise reduction process, and an electronic zoom process.

Explanation of symbols

1: Color separation prism, 2a-2c: CCD, 3a-3c: CDS & AGC, 4a-4c: A / D, 5a-5f: Multiplier, 6a-6c: Adder, 7: Field memory, 8: Memory control circuit 8a: memory input port, 8b, 8c: memory output port, 9: zoom interpolation circuit, 10: DSP, 11a, 11b: D / A, 12: adder, 13: connection terminal, 14: CCD drive signal generator , 15: field counter, 16: CPU.

Claims (1)

In a television camera that increases storage sensitivity, reduces noise, and enlarges an image, an image sensor that exposes incident light for one or more fields and outputs a video signal, a digital converter that converts the video signal to digital data, and a multiplier by changing the coefficients (value between 0 and 1), a first multiplier for multiplying the multiplication coefficient to the digital data, an adder for adding the output data and the third data of the first multiplier a memory for recording the output data of the adder, a value obtained by subtracting a multiplication factor (value between 0 and 1) of the first multiplier from 1 to 1 field delayed data from being recorded in the memory multiplier and comprises at least a second multiplier for outputting the third data, said recorded in the memory as a video signal to two-field delay data from the video data interpolation at the time of increasing the accumulation sensitivity Serial noise reduction and the image enlargement, television camera and performing by said arithmetic unit and read control of the memory.