JP2005229321A - Image pickup device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that an impression that an image appears stragglingly is made remarkable to degrade the image quality and S/N is degraded because three-dimensional noise reduction must be inevitably turned off, in the case of reduction of a shutter speed in a night vision mode. <P>SOLUTION: An LPF 15 has low pass characteristics of gradual attenuation from a low-band frequency near a DC to a frequency fs/4 and has frequency characteristics of increasing an output again in a high-band frequency region higher than the frequency fs/4. In the night vision mode, a microprocessor 18 selects gain 0/16 for a selection means 16 and selects gain 16/16 for a selection means 17. Then an image pickup signal which has such characteristics that at least a part of a high-band frequency component is more attenuated than a low-band frequency component, and has a gain increased in the low-band frequency is obtained from an adder 19. A night vision image can be picked up in a well-balanced condition with respect to frequency characteristics, S/N, and gain. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image pickup apparatus, and more particularly, a disc camera using a DV-type camera-integrated VTR, a DVD (Digital Versatile Disk) or the like as a recording medium. The present invention relates to an imaging apparatus such as a semiconductor camera using a semiconductor as a recording medium.

  In recent years, particularly in a camera-integrated VTR, there is a function equipped with a shooting mode (night vision mode) specialized for dark scenes, which is used for shooting a sleeping face of a child. Normally, when shooting a dark scene, an AGC (Auto Gain Control) circuit that adjusts the gain of the entire video signal compensates for underexposure by increasing the gain, but the higher the AGC gain, the higher the gain. This is accompanied by the side effect that S / N is deteriorated.

  In contrast, in the night vision mode, the AGC gain is not increased to avoid the S / N deterioration, but instead the shutter speed is slowed down so that light is accumulated in the image sensor for a long time, resulting in insufficient exposure. Therefore, a bright and relatively good S / N image is obtained even in the dark.

  By the way, the number of pixels of image pickup devices such as CCDs in recent image pickup devices has increased remarkably. For example, digital still cameras with more than 5 million pixels have been commercialized. Also, ones with 1 to 2 million pixels have been used.

  In order to obtain a standard television system video signal such as the current NTSC system, it is not necessary to have such a large number of pixels. However, in recent imaging apparatuses, in addition to a moving image medium such as a magnetic tape, for still image shooting. In many cases, a semiconductor memory card can be mounted. In such an image pickup apparatus, a multi-pixel image pickup device is provided, which is comparable to a dedicated digital camera as a still image camera for recording in a semiconductor memory. The feature of the device is that it can achieve high image quality.

  In addition, when such a multi-pixel image sensor is provided to obtain a standard television video signal, the output is finally reduced to a size suitable for the standard television video signal. Since the resizing by resampling performed at the time of image reduction is equivalent to averaging with surrounding pixels as a result, noise can be reduced, so that a multi-pixel image sensor is a standard television video signal. There is also an advantage that an image having a good S / N can be obtained.

  FIG. 4 is a block diagram showing an example of a conventional imaging apparatus. This conventional image pickup apparatus is a camera-integrated recording apparatus, which captures an external subject image by an image pickup unit 1 including an image pickup element such as a CCD (Charge Coupled Device) and its peripheral circuits, a lens system, an optical filter, and the like. Take an image. An imaging signal obtained by imaging by the imaging unit 1 is converted into a digital signal by the A / D converter 2. Now, it is assumed that this digital signal is quantized to an amplitude of 12 bits.

  Here, in a single-panel video camera, a color filter in which transmitted colors are arranged in a mosaic pattern is arranged in front of the imaging device, and the imaging signal is shown in FIGS. 5A and 5B for each field, for example. The component that represents is output. In FIGS. 5A and 5B, G, Cy, Ye, and Mg represent color components received by the light transmitted by the color filters that transmit green, cyan, yellow, and magenta, respectively. In this signal, two sets of mixed color components of two colors are alternately output for each sample (for details, see Non-Patent Document 1, for example).

  Such an image pickup signal is subjected to a known YC conversion process in the YC conversion unit 3 and is made into a set of a total of three signals including a luminance signal Y and two types of color signals constituting a color signal. When recording as a still image, the first recording processing unit 4 performs compression processing such as JPEG using the set of these three signals, and records the first recording medium 5 such as a semiconductor memory card.

  On the other hand, when recording as a moving image of a standard television signal, the set of three signals output from the YC converter 3 is band-limited by a low-pass filter (LPF) 7 in the resizing circuit 6 and then resized. Re-sampling by the re-sampling unit 8 in the circuit 6 results in resizing (reducing the number of pixels) to a signal with the number of samples having a specification that matches the standard television signal.

  The video signal output from the resizing circuit 6 is processed in a form suitable for recording a moving image such as a DV format or a DVD-RAM format in the second recording processing unit 9, and then the DV tape, DVD-RAM or the like. Are recorded on the second recording medium 10. The operation of the apparatus is performed by a user interface 11 such as a button, and the control of each block is performed by the microprocessor 12 as a response to the operation.

  In addition, an imaging apparatus for securing a visual field in an extremely low illuminance environment and improving recognition of a visually recognized object has been conventionally known (see, for example, Patent Document 1). This conventional imaging device is configured to improve visibility by reducing the luminance level by detecting the high-luminance portion in one screen and lowering the level of only the low-frequency component at that location (decreasing the absolute value). is there.

“Driving method of color CCD and actual signal processing”, transistor technology, Japan, CQ Publishing Co., Ltd. JP 2001-268410 A

  However, in the conventional imaging apparatus shown in FIG. 4, when the shutter speed is slowed down to cope with underexposure, especially when the shutter speed is slowed down to 1/60 seconds, the dynamic resolution is lowered, and the feeling that the image appears flickering is remarkable. There is a problem that the image quality deteriorates. In addition, when the 3D noise reduction is turned on in the slow shutter state, the image quality is further deteriorated. Therefore, there is a problem that the S / N deteriorates because the 3D noise reduction has to be turned off. .

  On the other hand, in the conventional imaging device described in Patent Document 1, a high-luminance portion in one screen is detected and processing of the detected portion is performed. Although it is improved, it is not a process for underexposure of the entire screen, and a night vision image of the entire screen cannot be captured with good balance in terms of frequency characteristics, S / N, and gain without using a slow shutter.

  The present invention has been made in view of the above points, and it is an object of the present invention to provide an imaging apparatus capable of capturing a night vision image in a balanced manner with respect to frequency characteristics, S / N, and gain without using a slow shutter as much as possible. And

  In order to achieve the above object, the present invention captures an object with an image sensor in either a normal mode for shooting a normal scene or a night vision mode provided for shooting a particularly dark scene. In the night-vision mode, the imaging device that attenuates at least a part of the high-frequency component of the imaging signal from the low-frequency component with respect to the input imaging signal obtained by the imaging device, A frequency selection unit that performs gain amplification of the component, and a second gain selected for the imaging signal processed by the frequency selection unit, and a first gain selected for the input imaging signal that has not been processed by the frequency selection unit Processing by the frequency selection means and the input image pickup signal to which the first and second gains are respectively given by the gain selection means It is obtained by a structure having an adding means for adding and outputting imaging signal.

  In the present invention, in the night vision mode, an image pickup signal having a characteristic in which at least a part of the high frequency component is attenuated more than the low frequency component and the gain in the low frequency component is enhanced is output from the adding means. be able to.

  According to the present invention, in an imaging device having a night vision mode, at least a part of a high frequency component is attenuated with respect to a low frequency component in the night vision mode, and the gain in the low frequency component is increased. Since the captured image signal is output, the image signal of the entire screen can be obtained with a well-balanced setting regarding frequency characteristics, S / N, and gain, instead of a specific signal portion within the screen, and it is as slow as possible. A captured image in a suitable night vision mode can be obtained without using a shutter.

  FIG. 1 shows a block diagram of an embodiment of an imaging apparatus according to the present invention. In the figure, the same components as those in FIG. 4 are denoted by the same reference numerals, and the description thereof is omitted. In FIG. 1, a digital image pickup signal output from the image pickup unit 1 and AD-converted by the A / D converter 2 is directly supplied to the first selection means (SEL1) 16, while a digital low-pass filter (LPF). After being band-limited by 15, it is supplied to the second selection means (SEL 2) 18.

  Here, as shown in FIG. 1, the LPF 15 is a so-called transversal type in which a plurality of delay means (denoted by τ), each of which is delayed by a unit time, are cascade-connected and gain is added to the tap output for addition. It is a filter. The characteristics of the LPF 15 which is this transversal filter are shown in FIG. In FIG. 6A and FIGS. 5B and 5C described later, fs represents a sampling frequency in the digital image pickup signal, and fs = 1 /, where τ is the same as the unit delay time of the delay means. τ relationship.

  The frequency characteristic of the LPF 15 shown in FIG. 2 (a) has a low-pass characteristic that gradually attenuates from a low-frequency near DC to a frequency fs / 4, and is output again in a high-frequency region above the frequency fs / 4. Is a frequency characteristic that increases. Here, in the imaging apparatus, the high-frequency component of the subject image input to the imaging unit 1 is attenuated at the stage of the light to be imaged by the optical filter provided in front of the imaging device, and this is electrically converted. When it is expressed in association with the signal stage, it is as shown by a broken line I in FIG. When this is combined with the frequency characteristics of the LPF 15 shown in FIG. 2A, it is as shown by a solid line II in FIG. 2B. The frequency characteristic indicated by the solid line II indicates a low-pass characteristic in the frequency region of the frequency fs / 4.

  Here, as shown in FIG. 1, the LPF 15 taps from the two delays of τ, as shown in FIG. 5, in the digital imaging signal considered here, 2 samples. This is because the same color component appears every time, so that the filtering process can be performed for each color component by using only this one LPF 15.

  In this LPF 15, as shown in FIG. 1, the signals from the three taps are added, but the gains of the first and third taps closer to the input are 1, and the gain of the center tap is 2. In order to set the gain at the input / output of the LPF 15 to 1 in the low frequency range, it is necessary to attenuate the output of the adder 21 in the LPF 15 to 1/4, but by intentionally not performing this attenuation, the input / output of the LPF 15 The gain at is increased to 4. Such an output signal of the LPF 15 is input to the second selection means (SEL2) 17.

  Here, in FIG. 1, the number of quantization bits at each part of the digital signal of each part is shown, and after AD conversion of the output image pickup signal of the image pickup unit 1, it is quantized to 12 bits. The number of bits of the output signal of the LPF 15 is 14 because the signal is amplified by the LPF 15.

  The first and second selection means 16 and 17 switch the output at the input and output from several set values and output both output signals of these selection means 16 and 17 by the adder 19. By adding, the ratio of the digital imaging output signal itself and the signal that has been filtered and amplified by the LPF 15 is determined.

  The gain selection by these selection means 16 and 17 is performed by a command from the microprocessor 18. An example of the selection is shown in FIG. In the example of FIG. 5A, in the normal mode, the selection means (SEL1) 16 outputs a gain of 16/16, that is, the digital imaging signal as it is with gain of 1, and the selection means (SEL2) 17 has a gain of 0/16. That is, the signal output from the LPF 15 is turned off.

  Therefore, in the normal mode, the signal output from the adder 19 is the digital imaging signal itself output from the A / D converter 2. Here, when the microprocessor 18 detects that the night-vision mode is selected from the user interface 11 by the user's operation, the microprocessor 18 instructs the selection means (SEL1) 16 as shown in FIG. A command is issued to select gain 16/16 and gain 16/16 to selection means (SEL2) 17.

  Therefore, in the night vision mode, the signal output from the adder 19 is the signal itself output from the LPF 15. Accordingly, the frequency characteristic of the signal output from the adder 19 is as shown by the solid line II in FIG. 2B, and the low-frequency gain near DC is enhanced by a factor of four compared to the normal time.

  As another example of the gain selection by the selection means 16 and 17, FIG. In this case, the normal mode is the same as the example of FIG. 3A, but in the night vision mode, the gain of the digital imaging signal itself and the output signal of the LPF 15 is 8/16. Therefore, when the gain shown in FIG. 3B is selected, the frequency characteristic of the output signal of the adder 19 is as shown by a solid line III in FIG.

  The frequency characteristic indicated by the solid line III is not as high as the frequency characteristic indicated by the solid line II in FIG. 2B, but the gain in the low band is increased and the band near fs / 4 is slightly higher than in the normal mode. It is a characteristic that suppresses it. This is suitable when it is desired to secure S / N and a high frequency component as much as possible on the assumption that there is a certain level of brightness even in the night vision mode.

  As described above, the selection of the gain in the selection means 16 and 17 is switched between the normal mode and the night vision mode, and in the output signal of the adder 19 on the output side of the selection means 16 and 17, the night vision mode is better. At the same time, at least a part of the high-frequency component is attenuated more than the low-frequency component, and the gain at the low-frequency is enhanced.

  Returning to FIG. 1 again, the quantization bit number of the output signal from the selection means 16 is 12 bits, but the quantization bit number of the output signal from the selection means 17 is 14 bits. The number of quantization bits of the 19 output signals is 14 bits, but the processing after the YC conversion unit 3 similar to the conventional processing is the same as the conventional processing when the recording signal and the final output signal are the same as the conventional processing. Preferably done in bits.

  Therefore, in this embodiment, in the 12-bit limiter 20, the upper 2 bits of the output signal of the adder 19 are deleted and limited to 12 bits. By doing so, the high level signal is saturated and the gradation of the bright part is lost, but such saturation occurs in the night vision mode in which the component whose gain is increased by the LPF 15 is added. It is not supposed to shoot a bright scene that is originally saturated, and since it is already understood that the user is a mode for shooting a dark scene from the mode name "Night Vision Mode", this is not a problem. . As described above, in the night vision mode, it is possible to shoot with emphasis on details of a dark scene.

  The processing after the YC conversion unit 3 is the same as that of the conventional imaging apparatus shown in FIG. 4, and the first recording processing unit 4 records a video signal related to a still image on the first recording medium 5 such as a semiconductor memory card. Then, the second recording processing unit 9 generates a moving image signal having a number of samples that matches the standard television signal after being resized (reduced in the number of pixels), and the second recording medium 10 such as a DV tape or a DVD-RAM. To be recorded.

  As described above, according to the present embodiment, it is possible to capture a night-vision image of the entire screen in a balanced manner with respect to frequency characteristics, S / N, and gain without using a slow shutter as much as possible. This embodiment is particularly suitable when a multi-pixel CCD is used as an image sensor. In that case, the resizing process is performed to match the number of pixels of the imaging signal to that of the standard signal, so that there is an S / N improvement effect, and there is a certain margin of S / N. This is because, if at least a part of the signal is attenuated, it is allowed to increase the gain as much as four times.

  Note that the present invention is not limited to the above embodiment. For example, the selection of the gains in the selection means 16 and 17 may be performed in the night vision mode with respect to the low frequency component in the output signal of the adder 19. As long as at least a part of the high frequency component is attenuated, it is only necessary to increase the gain in the low frequency range, and it is limited to the specific selection shown in FIGS. 3 (a) and 3 (b). It is not something that can be done.

  Further, the specific selection of the gains in the selection means 16 and 17 regarding the night vision mode may be determined and set as a specific one at the time of designing the imaging apparatus, and may be set by user operation. It may be possible to select from several combinations, and further, it may be selected automatically from several combinations according to conditions such as brightness at the time of photographing detected by the sensor. Good.

It is a block diagram of one embodiment of the present invention. It is a frequency characteristic figure of the output signal of LPF or an adder which is the principal part of FIG. It is explanatory drawing of the gain which the selection means in FIG. 1 selects in normal mode and night vision mode. It is a block diagram of an example of the prior art. It is explanatory drawing of an example of an imaging signal.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image pick-up part 3 YC conversion part 4 1st recording process part 5 1st recording medium 6 Resize circuit 9 2nd recording process part 10 2nd recording medium 15 Low pass filter (LPF) by a transversal filter
16, 17 Selection means 18 Microprocessor 19, 21 Adder 20 12-bit limiter

Claims (1)

In an imaging device that captures an image of an object with an image sensor in either a normal mode for capturing a normal scene or a night vision mode provided for capturing a particularly dark scene,
In the night vision mode, for the input image signal obtained by the image sensor, at least a part of the high frequency component of the image signal is attenuated from the low frequency component, and the low frequency component Frequency selection means for performing gain amplification;
A gain selection unit that applies the selected first gain to the input imaging signal that has not been processed by the frequency selection unit and also applies the selected second gain to the imaging signal that has been processed by the frequency selection unit;
And an addition means for adding and outputting the input imaging signal to which the first and second gains are respectively given by the gain selection means and the imaging signal after being processed by the frequency selection means, and the night vision mode. In some cases, the adding means outputs an image pickup signal having a characteristic in which at least a part of the high frequency component is attenuated more than the low frequency component and the gain in the low frequency component is increased. A characteristic imaging apparatus.

Images