JP2004320733A - Method for processing information signal and video camera - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that the dynamic range for detecting infinitesimal shaking amplitudes is not enough in a recent video camera of high image quality, and a wide dynamic range is preferably used for an A-D converter, but it is not practical. <P>SOLUTION: A sensor signal of usual gain amplified by an amplifier 39 (40) from the output sensor signal of a shaking sensor 37 (38) and a sensor signal of magnification gain amplified by amplifiers 39 and 44 (40 and 45) are generated, alternately switched by an SW 46, converted into a digital signal by the A-D converter 41 (42), and supplied to a microcomputer 43. The sensor signal of the usual gain is bit extended by a bit extension unit 43b, and stored in a memory 54, and the sensor signal of the magnification gain is stored in a memory 55 by the microcomputer 43. Only when the sensor signal of the bit extended usual gain is less than predetermined, a shake correction signal is generated by using the sensor signal of the magnification gain of the memory 55. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information signal processing method and a video camera, and more particularly to an information signal processing method for processing a sensor signal obtained by detecting camera shake using an angular velocity sensor in a transmission system having a finite dynamic range, and a video camera equipped with a camera shake correction device. .

  Conventionally, in order to prevent image shake due to camera shake that is likely to occur during handheld shooting of a video camera, shake information of the video camera is detected by shake detection means, and the shake is optically or electronically canceled according to the detection result. Various types of camera shake correction devices that realize camera shake correction by doing so have been proposed.

  The camera shake correction device includes, for example, an active prism provided in front of a CCD image sensor for correcting an optical axis of imaging light from a subject, an angular velocity sensor for detecting vertical and horizontal angular velocities of a video camera, and an angular velocity sensor And a controller for driving the active prism based on the output angular velocity data.

  The control unit has a high-pass filter and an integration circuit, removes a DC component from the angular velocity data by the high-pass filter, and converts the angular velocity data indicating the angular velocity into angle data indicating the angle of camera shake by the integration circuit. It is converted and supplied to the active prism. The active prism corrects the optical axis of the imaging light in accordance with the supplied angle data, and performs camera shake correction so that light from the same position of the subject is received at the same position of the CCD image sensor.

  However, the conventional video camera equipped with the above-described image stabilization apparatus has an active prism in a direction opposite to the panning or tilting direction because the image stabilization is performed by the image stabilization apparatus even when panning or tilting is performed. Is corrected, so-called swing-back phenomenon occurs, and there is a problem that the position of the subject in the captured image does not change despite performing panning and tilting.

  Therefore, by switching the circuit characteristics of the above-described high-pass filter, discriminating between camera shake and panning or tilting, and performing camera shake correction according to the discrimination result, it is possible to improve the accuracy of camera shake correction. A camera has been conventionally proposed (for example, see Patent Document 1).

Japanese Patent Application Laid-Open No. 7-288834 (FIG. 1)

  By the way, in recent years, the image quality of video cameras has been improved, and the accuracy required for the camera shake correction function itself has also been increased. In particular, it has become necessary to be able to surely correct even a small-amplitude camera shake, and a wide dynamic range is required for the camera shake detection means to cope with this.

  However, in the conventional video camera described in Patent Literature 1, at the time of camera shake correction, an angular velocity sensor detects vertical and horizontal angular velocities and supplies them to a 10-bit A / D converter via a filter and an amplifier circuit. Thus, the angular velocity detection data is formed, and furthermore, a camera shake detection signal is generated from the angular velocity detection data. Therefore, the dynamic range is not sufficient for the detection of a small amplitude camera shake. To cope with this, it is sufficient to use an A / D converter having a wide dynamic range, but this is not practical, and there are problems such as an increase in the size of the camera shake correction device and a reduction in processing speed.

  The present invention has been made in view of the above points, and converts a sensor signal output from an angular velocity sensor into a plurality of sensor signals to which gains different from each other are provided, and a value of a sampling point of the plurality of sensor signals defines a dynamic range. By selecting a value of the sampling point of the sensor signal having the optimum gain or interpolating from the values of the plurality of sampling points depending on whether or not the value exceeds the maximum value, it is possible to increase the value of the camera shake correction device without increasing the circuit scale. It is an object of the present invention to provide an information signal processing method and a video camera capable of securing a dynamic range.

  In order to achieve the above object, an information signal processing method according to a first aspect of the present invention is an information signal processing method for transmitting an input analog signal through a transmission system having a finite dynamic range. A first step of converting the transmitted analog signals into a plurality of analog signals to which a transmission system is transmitted, and a step of separately sampling the transmitted plurality of analog signals with a predetermined sampling clock to generate a plurality of digital signals. Step 2 and when the values of the sampling points of the plurality of digital signals do not exceed the dynamic range, select the values of the sampling points of the digital signal to which the maximum gain is applied, and select one or more digital signals. When the value of the first sampling point exceeds the dynamic range, the digital value not exceeding the dynamic range. Selecting a value interpolated using the value of a second sampling point adjacent to one or both sides of the first sampling point of the digital signal to which the largest gain has been given among the signals, The method is characterized in that an analog signal is output in which the values of the sampling points selected in the third step are synthesized in time series.

  According to the present invention, when the values of the sampling points of a plurality of digital signals having different gains from each other do not exceed the dynamic range, the values of the sampling points of the digital signal having the maximum gain are selected, and one or more of the digital signals are selected. When the value of the first sampling point exceeds the dynamic range, the second signal adjacent to one or both sides of the first sampling point of the digital signal having the largest gain among the digital signals not exceeding the dynamic range. Since an interpolated value is selected using the value of the sampling point, a small-amplitude analog signal can be transmitted as a digital signal having as large a gain as possible in a system having a finite dynamic range.

  In order to achieve the above object, an information signal processing method according to a second invention is an information signal processing method for transmitting an input analog signal through a transmission system having a finite dynamic range. A first analog signal obtained by enlarging a signal portion having a small amplitude equal to or less than a value and a second analog signal representing the amplitude of a signal portion having a larger amplitude than a predetermined value, and transmitting the signal through a transmission system; A step of separately sampling the transmitted first and second analog signals with a predetermined sampling clock to generate first and second digital signals, and a sampling point of the first digital signal. If the value of the first digital signal does not exceed the dynamic range, the value of the sampling point is selected, and the value of the first sampling point of the first digital signal exceeds the dynamic range. The interpolation value larger than the dynamic range calculated by a predetermined arithmetic expression using the values of the second sampling points adjacent to both sides of the first sampling point of the second digital signal. And outputting a sensor signal in which the values of the sampling points selected in the third step are synthesized in time series.

  According to the present invention, of the first digital signal obtained by enlarging the signal portion having a small amplitude equal to or smaller than a predetermined value from the input analog signal and the first digital signal representing the amplitude of the signal portion having a larger amplitude than the predetermined value, When the value of the sampling point of the digital signal does not exceed the dynamic range, the value of the sampling point is selected. When the value of the first sampling point of the first digital signal exceeds the dynamic range, An interpolation value larger than a dynamic range calculated by a predetermined arithmetic expression using values of second sampling points adjacent to both sides of the first sampling point of the second digital signal is selected. In a system having a finite dynamic range, a small-amplitude sensor signal can be transmitted as a first digital signal having as large a gain as possible.

  Further, in order to achieve the above object, a video camera according to a third aspect of the present invention temporarily stores a video signal obtained by imaging with an image sensor in a video signal memory and a sensor signal obtained by an angular velocity sensor. Is transmitted in a transmission system having a finite dynamic range, and based on the camera shake correction signal obtained from the transmitted sensor signal, the readout position of the video signal memory is variably controlled, so that the video signal corrected from the video signal memory is A conversion means for converting a sensor signal obtained by an angular velocity sensor into a plurality of sensor signals provided with different gains, and separately sampling the plurality of sensor signals with a predetermined sampling clock. A / D conversion means for generating a plurality of digital sensor signals, and a sampler for the plurality of digital sensor signals When the value of the sampling point does not exceed the dynamic range, the value of the sampling point of the digital sensor signal having the maximum gain is selected, and the value of the first sampling point of one or more digital sensor signals determines the dynamic range. If it exceeds, the interpolation is performed using the value of the second sampling point adjacent to one or both sides of the first sampling point of the digital sensor signal having the largest gain among the digital sensor signals not exceeding the dynamic range. It is characterized by comprising a selection means for selecting a value, and a camera shake correction signal generation means for generating a camera shake correction signal based on a time-series synthesized signal of the values of the sampling points selected by the selection means.

  According to the present invention, when the values of the sampling points of the plurality of digital sensor signals having different gains do not exceed the dynamic range, the value of the sampling point of the digital sensor signal having the maximum gain is selected, and one or more digital values are selected. When the value of the first sampling point of the sensor signal exceeds the dynamic range, one or both sides of the first sampling point of the digital sensor signal having the largest gain among the digital sensor signals not exceeding the dynamic range. Since the interpolated value is selected using the value of the adjacent second sampling point, the small-amplitude sensor signal is converted into a digital sensor signal having as large a gain as possible and transmitted, and camera shake correction is performed based on the digital sensor signal. A signal can be generated.

  In order to achieve the above object, a video camera according to a fourth aspect of the present invention includes a first sensor signal obtained by enlarging a signal portion having a small amplitude equal to or smaller than a predetermined value from a sensor signal obtained by an angular velocity sensor, and Converting means for converting the signal into a second sensor signal representing the amplitude of the large-amplitude signal portion; and first and second digital sensor signals obtained by separately sampling the first and second sensor signals with a predetermined sampling clock. A / D conversion means for generating the first digital sensor signal, and when the value of the sampling point of the first digital sensor signal does not exceed the dynamic range, the value of the sampling point is selected. When the value of the sampling point exceeds the dynamic range, the second digital sensor signal of the second sampling point adjacent to both sides of the first sampling point is output. Selecting means for selecting an interpolated value larger than the dynamic range obtained by calculating a predetermined arithmetic expression using, and a camera shake correction signal based on a time-series synthesized signal of sampling point values selected by the selecting means. And a camera shake correction signal generating means for generating. According to the present invention, in a system having a finite dynamic range, a small-amplitude sensor signal can be transmitted as a first digital sensor signal having a gain as large as possible.

  Further, in order to achieve the above object, the video camera of the present invention converts an incoming light into a video signal by an image sensor, temporarily stores the video signal in a video signal memory, and obtains an angular velocity obtained by using an angular velocity sensor. In a video camera that calculates a read position of a video signal memory from a signal and changes a read position of a video signal in the video signal memory to realize camera shake correction, an output signal from an angular velocity sensor is amplified by a first gain value. And a first signal amplifying means for outputting a first amplified signal, and amplifying the output signal from the angular velocity sensor with a second gain value larger than the first gain value, and outputting a second amplified signal. Second signal amplifying means, selecting means for switching the first and second amplified signals at a cycle twice as long as the video cycle for driving the video signal memory, and first and second amplified signals selected by the selecting means To An A / D converter for sampling, a bit extension means for correcting the gain of the first digital signal obtained by sampling the first amplified signal by the A / D converter by bit extension, and a second signal output from the bit extension means. A first memory for storing one digital signal, a second memory for storing a second digital signal obtained by sampling a second amplified signal by an A / D converter, and a first and a second memory. Memory output selection means for selecting one of the first and second digital signals respectively output from the memory, and a value of the gain-corrected first digital signal output from the first memory reaching a predetermined value Is detected, the memory output selection means selects the first digital signal output from the first memory, and the value of the gain-corrected first digital signal becomes a predetermined value. When it is full, the memory output means controls the memory output selection means so as to select the second digital signal output from the second memory, and the second output signal selected and output by the memory output selection means. A camera shake correction signal generating means for generating a camera shake correction signal indicating a read position of the video signal memory based on the first or second digital signal, wherein the camera shake correction is performed by changing the read position of the video signal memory based on the camera shake correction signal. Is realized.

  According to the information signal processing method of the present invention, in a system having a finite dynamic range, a small-amplitude sensor signal can be transmitted as a digital sensor signal having a gain as large as possible. As a result, a sensor signal having a small amplitude can be transmitted with high accuracy using a signal having a relatively small dynamic range.

  According to the video camera of the present invention, among the sensor signals obtained by detecting the camera shake by the angular velocity sensor, the sensor signal having a small amplitude is transmitted as a digital sensor signal having a gain as large as possible, and the camera shake is performed based on the digital sensor signal. Since the correction signal can be generated, a small-amplitude sensor signal can be transmitted with high accuracy by using an A / D converter having a relatively small dynamic range as an A / D conversion means for generating a digital sensor signal. The camera shake correction can be performed with higher accuracy than before without increasing the image quality.

  Furthermore, according to the present invention, a high number of bits is effective even in the case of a small-amplitude camera shake, so that an error due to arithmetic processing by the microcomputer is small, and the camera shake correction amount can be output with high accuracy, and the camera shake correction accuracy A video camera capable of improving image quality can be provided.

  Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1A is a schematic block diagram of an embodiment of a video camera according to the present invention, and FIG. 1B is a block diagram of an embodiment of a main part of FIG.

  First, the processing of an incoming video will be described sequentially with reference to FIG. The imaging subject light coming from a lens (not shown) is imaged on a CCD (Charge Coupled Device) 31 which is a solid-state imaging device, photoelectrically converted into an imaging signal, and then digitally converted by an A / D conversion unit 32 to be well known. After being shaped into a video signal by the signal processing circuit 33, the video signal is supplied to the memory 34 and stored.

  On the other hand, in the camera shake detection portion, sensor signals output from a camera shake sensor 37 in the yaw (Yaw) direction (or yawing direction) and a camera shake sensor 38 in the pitch (Pitch) direction, which are angular velocity sensors, are respectively amplified. After being separately amplified by 39 and 40, the signal is directly supplied to a switch (SW) 46, and further separately amplified by amplifiers 44 and 45 and supplied to SW 46. The sensor signal in the yaw direction and the sensor signal in the pitch direction selected by the SW 46 are supplied to A / D converters 41 and 42 and separately converted into digital signals. Supplied.

  The microcomputer 43 performs camera shake sensor signal processing, which will be described later, generates a camera shake correction signal, and changes the readout position of the memory 34 to perform camera shake correction. The digital video signal read from the memory 34 is converted into an analog signal by the D / A converter 35 and then supplied to the monitor 36 to be displayed as a good subject image without camera shake.

  Next, processing of a camera shake sensor signal will be described with reference to FIG. FIG. 2A shows the waveform of the camera shake sensor 37 or 38, and FIG. 2B shows the output waveform of the amplifier 39 or 40 obtained at the normal gain. If there is a camera shake, a sensor signal of a level corresponding to the magnitude of the camera shake is extracted from both or one of the camera shake sensors 37 and 38 according to the direction of the camera shake. Since the input signals of the amplifiers 44 and 45 are the output signals of the amplifiers 39 and 40, the output signal waveforms of the amplifiers 44 and 45 are enlarged as shown in FIG. It becomes a gain waveform.

  The two sensor signals, the sensor signal amplified by the normal gain and the sensor signal amplified by the enlarged gain, are switched by the SW 46 and are digitally converted by the A / D converters 41 and 42. It is supplied to the microcomputer 43 in a D range (dynamic range) 1. Inside the microcomputer 43, a synthesized signal of the D range 2 indicated by bit expansion described later is obtained as shown in FIG. The small amplitude portion I in the normal gain sensor signal shown in FIG. 2B is enlarged from the magnitude of the normal gain by the amplifier 44 or 45 to the magnitude indicated by II in the composite signal shown in FIG. The bit accuracy is improved. Based on the synthesized signal, a camera shake correction signal is generated.

  Next, a method of synthesizing two sensor signals having different gains (gains) will be described with reference to the block diagram of FIG. 1B, the time chart of FIG. 3, and the flowchart of FIG. FIG. 1B is a detailed block diagram of a gain signal combining portion inside the microcomputer 43. For convenience of explanation, FIG. 1B shows a processing section 43a of only a sensor signal detected by the yaw (Yaw) direction camera shake sensor 37. I have.

  The digital sensor signal a in the yaw direction supplied from the A / D converter 41 in FIG. 1A converts the output signal of the amplifier 39 during the connection period of the terminal (0) of the SW 46. This is sensor signal data of normal gain obtained, and sensor signal data of enlarged gain obtained by A / D converting the output signal of the amplifier 44 during the connection period of the terminal (1) of the SW 46.

  The sensor signal data of the normal gain is digitally amplified (gain corrected) by the gain of the amplifier 44 in the bit expansion unit 43b of FIG. Here, if the gain of the amplifier 44 is “2”, the bit expansion unit 43b expands the sensor signal data of the normal gain by one bit and sets the LSB thereof to the same level as the sensor signal data of the expansion gain. Set to 0.

  Further, the sensor signal data of the above-described enlarged gain is directly supplied to the memory 55 and stored. The switching timing of the above-described SW 46 and the storage timing in the memories 54 and 55 are controlled by a switch control (SWCTL) unit 43f.

  FIG. 3 is a time chart showing the operation of acquiring the A / D conversion data of the gain signal. As shown in FIG. 3A, the SWCTL unit 43f converts the A / D conversion and data acquisition pulse into two of the conventional acquisition pulse. The switching pulse d of the SW 46 shown in FIG. 5B is generated at a frequency twice as long (１ ／ times the cycle), synchronized with the data capturing pulse, and with a cycle twice the data capturing pulse. To the switch 46 to control the switching.

  The above-described A / D conversion and data capture pulse is supplied to the A / D converter 41 as a sampling pulse and is also supplied to the memories 54 and 55 as a write pulse. The switching pulse d is supplied to the SW 46 to control the switching of the SW 46, and is also supplied to the memories 54 and 55 so that only one of the memories can perform the write operation, and the memory capable of the write operation is alternately switched. Here, the memory 54 is controlled to be in a write-enabled state during the low level period of the switching pulse d, and the memory 55 is controlled to be in a write-enabled state during the high level period.

  The SW 46 is connected and controlled so as to select the input signal of the terminal (0) during the low level period of the switching pulse d, whereby the sensor signal data of normal gain from the amplifier 39 via the terminal (0) is transmitted to the bit extension unit 43b. After the bit is extended (gain correction) in the memory 54, the data is stored in the memory 54 based on the data capture pulse shown in FIG. The SW 46 is connected and controlled so as to select the input signal of the terminal (1) during the high level period of the switching pulse d, whereby the sensor signal data of the expanded gain from the amplifier 44 via the terminal (1) becomes high. The data is stored in the memory 55 based on the data capture pulse shown in FIG.

  Accordingly, the data storage states of the memories 54 and 55 are as schematically shown in FIGS. 3 (c) and 3 (d). Thereafter, by selectively combining the amplified sensor signal data stored in the memories 54 and 55 with the SW 43c as described later, a correction use value c for use in camera shake correction schematically shown in FIG. can get.

  Next, the operation of storing data in the memories 54 and 55 and the operation of selectively combining sensor signals from the stored data will be described in more detail with reference to the flowchart shown in FIG. First, an operation of capturing sensor signal data will be described with reference to FIG. When a sensor signal (angular velocity signal) output from the SW 46 shown in FIG. 1A is input to the A / D converter 41 (step S50), the A / D converter 41 converts the sensor signal into an A / D signal. It is converted into sensor signal data (step S51).

  Next, it is determined whether or not the SW 46 is connected to the terminal (1) (step S52), and the sensor signal data from the A / D converter 41 is stored in the memory 55 during the period when the SW 46 is connected. (Step S54). On the other hand, when the SW 46 is connected to the terminal (0), the bit is expanded by the same gain as that of the amplifier 44 (or 45) by gain correction (step S53), and then stored in the memory 54 (step S55). Thus, the capture of the sensor signal data is completed (step S56).

  Next, an operation of selectively synthesizing sensor signal data and generating a camera shake correction signal will be described with reference to FIG. When the capture of the sensor signal data into the memories 54 and 55 is completed according to the procedure shown in FIG. 4A, generation of a camera shake correction signal is started (step S60).

  In the data control unit 43e shown in FIG. 1B, a reference value Com for determining the magnitude of camera shake is set in advance (step S61). The default value of the reference value Com is a value corresponding to the maximum value of the D range of the amplifiers 44 and 45, and is adjustable. The data control unit 43e first reads the gain-corrected (bit-extended) sensor signal data stored in the memory 54, and if the gain-corrected sensor signal data is larger than the reference value Com and It is determined whether or not the clock is continuous for more than the clock (step S62). The reason for determining whether two clocks, that is, two or more clocks having the same cycle as the SW46 switching pulse, is continuous is to prevent erroneous detection due to noise.

  If the data control unit 43e determines that the sensor signal data subjected to gain correction is larger than the reference value Com and continues for two or more clocks (Yes in step S62), the sensor signal of the enlarged gain stored in the memory 55 Since the data is saturated, the switch (SW) 43c is switched and controlled so that the gain-corrected sensor signal data stored in the memory 54 is read from the memory 54 and supplied to the camera shake correction signal generation unit 43d ( Step S63).

  On the other hand, when the gain-corrected sensor signal data stored in the memory 54 is equal to or smaller than the reference value Com (No in step S62), the sensor signal data of the enlarged gain stored in the memory 55 is not saturated, and Since the small amplitude portion is within the D range 1 as shown in FIG. 2 (c), the data control unit 43e reads out the sensor signal data of the enlarged gain stored in the memory 55 from the memory 55 and reads the camera shake correction signal. The switch (SW) 43c is controlled so as to be supplied to the generating unit 43d (step S64).

  When the gain-corrected sensor signal data stored in the memory 54 is equal to or less than the reference value Com, the gain-corrected sensor signal data is not selected without selecting the gain-corrected sensor signal data. Compared to the sensor signal data in which the gain is corrected by a simple bit extension of the digital signal by the unit 43b, the sensor signal of the expanded gain generated by A / D conversion after being amplified by the two amplifiers 39 and 44 This is because the data is data with higher resolution that is more faithful to the level of the small amplitude portion. Therefore, even if the same A / D converter as the conventional one, for example, a 10-bit A / D converter is used as the A / D converter 41, In addition, it is possible to perform more accurate camera shake correction for a small amplitude portion.

  As described above, the magnitude of the camera shake is determined in accordance with the comparison result between the gain-corrected normal gain sensor signal data stored in the memory 54 and the reference value Com, and is stored in the memory 54 or the memory 55. By determining which of the sensor signal data to use and reading it out, a time-series synthesized signal of D range 2 shown in FIG. 2D is obtained from the SW 43c. This composite signal corresponds to the correction use value c described above.

  Thereafter, in the camera shake correction signal generation unit 43d in FIG. 1B, DC (direct current) component removal is performed on the input sensor signal data (synthesized signal) by a high-pass filter (HPF) (step S65), and the HPF is used. The pan / tilt component is removed (step S66), the integration process by a low-pass filter (LPF) (step S67) is sequentially performed, and finally, the zoom gain process for correcting the zoom magnification (step S68) is performed.

  The camera shake correction signal generated by the camera shake correction signal generation unit 43d in this way is output to the memory 34 (Step S69). The read position of the memory 34 is controlled by the camera shake correction signal, and a video signal in which the camera shake in the yaw (Yaw) direction has been corrected is read from the memory 34.

  The processing unit 43g having the same configuration as the processing unit 43a in the yaw (Yaw) direction also applies to the digital sensor signal b in the pitch (Pitch) direction supplied from the A / D conversion unit 42 in FIG. Then, the same processing as described above is performed in parallel, and the video signal in which the camera shake in the pitch (Pitch) direction has been corrected is read from the memory 34.

  Next, another embodiment of the present invention will be described. FIG. 5A shows only the output of the amplifier 39 (or the amplifier 40, hereinafter referred to as the yaw direction sensor 37) shown in FIG. 1A, but the same applies to the output of the pitch direction sensor 38. 5) shows the output signal waveform of the amplifier 44. FIG. Since these signal waveforms are switched by the SW 46, the output signal waveform of the SW 46 is as shown in FIG.

  Here, assuming that the limit of the dynamic range of the A / D converter 41 is TH, the large amplitude signal is saturated at TH, so that the signal waveform as shown in FIG. You. The state of the sampling is shown in FIG. .., P1, p2,... Represent sampling points for the output signal waveform of the amplifier 39, and q1, q2,. The levels of these sampling points are used as digital signals.

  FIG. 5F shows digital information obtained by performing an interpolation process after the A / D conversion by the A / D conversion unit 41 as an analog waveform. When the output signal of the amplifier 44 is less than the limit (saturation level) TH of the dynamic range of the A / D converter 41, the information of the sampling points q1, q2,. Is greater than the dynamic range limit TH of the A / D converter 41, that is, when the value of the sampling point q * (* is a natural number) is TH, the sampling point p1 of the output signal of the amplifier 39 is , P2,..., Are interpolated.

  The interpolation method is, for example, as follows.

(First interpolation method)
In the first interpolation method, information on the sampling point of the output signal of the amplifier 39 adjacent to the sampling point of the output signal of the amplifier 44 to be interpolated (for example, in FIG. 5E, p5 or p6 for q5). The value doubled (after A / D conversion, it is digitized and expressed in a binary system, so it is only necessary to perform a bit shift) and interpolated as shown by r5 in FIG. 5 (f).

  When interpolating with a signal at a later point in time, a buffer is used so that processing can be performed using the signal coming later. Specifically, the sampling points p1, p2,... And q1, q2,. The signals shown in FIGS. 6A to 6E correspond to the signals shown in FIGS. 3A to 3E, and the correction use value is, as shown in FIG. When the value of * (* is a natural number) is TH, interpolation values r5 to r8 obtained by doubling the information of the sampling point p * of the output signal of the amplifier 39 are used. These values r5 to r8 are the values shown in FIG.

  In this case, the interpolated value has half the resolution in the amplitude direction, but has the advantage that it can be realized by a simple process of bit shifting. This decrease in resolution occurs at the time of large amplitude.For example, in a video camera for high-definition video shooting, correction of a large amplitude camera shake is to be performed at the same level as before, and since the image is high-definition, In particular, if the correction is made with high precision for small amplitude camera shakes, large amplitude camera shakes occur relatively infrequently, even if the user intends to hold the camera firmly. As a high-definition image is photographed, since the camera shake due to the minute shake is constant, even if the minute camera shake is corrected with high accuracy, it is a great improvement in the image obtained.

  In this example, the switching output by the SW is substituted by the value obtained from the value before or after the point to be interpolated in the sampled signal, and therefore, the information at one time before or after one point in time is substituted. Become. This does not substantially matter if the change in the signal of the camera shake detection output is sufficiently slow with respect to the sampling period.

(Second interpolation method)
In the second interpolation method, information on the sampling points (for example, p5 and p6) of the output signals of the amplifier 39 on both sides of the sampling point (for example, q5) of the output signal of the amplifier 44 to be interpolated is added as r5. Interpolate. In this case, since the interpolation is performed based on the information before and after the point to be interpolated, the information at the time point q5 is predicted. According to this method, for example, when the value of p5 is "6" and the value of p6 is "7", the added value is "13", which is a value obtained by predictive interpolation, but similar to the first interpolation method. There is no reduction in resolution.

  .. And q1, q2,... In FIG. 5 are specifically shown in FIG. 7 (a) to FIG. 7 (e). The signals shown in FIGS. 7A to 7E correspond to the signals shown in FIGS. 3A to 3E, and the correction use value is, as shown in FIG. When the value of * (* is a natural number) is TH, interpolation values r5 to r8 obtained by adding information on the sampling points of the output signals of the amplifiers 39 on both sides of the sampling point are used.

  As described above, according to the first and second interpolation methods, a small-amplitude signal is input to the A / D converters 41 and 42 with a large gain. For a large-amplitude signal that saturates the dynamic range, a signal with a smaller gain is arranged at an adjacent sampling point and interpolated from the signal, thereby achieving highly accurate correction for a small-amplitude camera shake. At the same time, it is possible to drive the correcting means by an amount corresponding to the large amplitude to the correction means by using the signal generated by interpolation even for a large amplitude camera shake.

  In the first interpolation method, the signal to be interpolated is a signal at a time point shifted by the sampling period from the point to be interpolated, but the interpolated signal and the interpolated signal are shifted so as not to cause this time lag. In order to adjust the time, a delay circuit for a sampling period may be inserted in the output of either the amplifier 39 or the amplifier 44 before inputting to the SW 46.

  In addition, when interpolating with information created from the previous sampling point (specifically, for example, in FIG. 5E, when interpolating the signal at the time point of q5 by doubling the signal of p5), When a delay circuit 47 (and 48) for one sample period is inserted into the amplifier 44 (and 45) side as shown in FIG. 8, and interpolation is performed using information generated from a subsequent sampling point (specifically, for example, FIG. In (e), when the signal at the time point of q5 is interpolated by doubling the signal of p6), as shown in FIG. 9, the delay circuit 49 (and 40) for one sample period is provided to the amplifier 39 (and 40) side. 50) may be inserted.

  In the above embodiment, two amplifiers 39 and 44 (or amplifiers 40 and 45) are prepared to generate two sensor signals having different gains from the output sensor signal of one sensor 37 (or 38). However, in order to obtain a waveform as shown in FIG. 5C, the output sensor signal of one sensor 37 (or 38) may be amplitude-modulated (AM-modulated) to obtain two sensor signals.

  That is, as shown in FIG. 10, the sensor signals output from the sensors 37 and 38 are amplitude-modulated by the AM modulators 101 and 102 and supplied to the A / D converters 41 and 42. The switching pulse d of the SW 46 is supplied to the AM modulators 101 and 102 as a carrier. As a result, the AM modulators 101 and 102 output a waveform in which two sensor signals having different gains are time-sequentially synthesized at the switching pulse period as an AM modulated wave. Here, the sensor output may be appropriately amplified by a preamplifier or the like provided after the sensors 37 and 38.

  Further, in the above-described embodiment, a description has been given of sequentially sampling signals of two different gains. However, the present invention is not limited to the signal of two gains. For example, as shown in FIG. And 203, the signals of three gains may be switched by the SW 206. Here, assuming that interpolation is performed using a later sampling point, the time points of the original signals sampled by the delay circuits 204 and 205 are aligned at the time of interpolation as in FIG.

  For example, assuming that the gain of the amplifier 201 that first amplifies the output sensor signal of the camera shake sensor is “1”, of the amplifiers 202 and 203 that amplify the output sensor signal of the amplifier 201, respectively, the gain of the amplifier 202 is “2”; The gain of the amplifier 203 is set to “4”. FIG. 12 shows a diagram corresponding to a combination of FIGS. 5E and 5F for the output sensor signals of the amplifiers 201 to 203 having these three gains.

  In FIG. 12, the time adjustment by the delay circuits 204 and 205 is also reflected. When the information at the sample point of the maximum gain does not reach the saturation level TH of the A / D converter 41 (or 42) (for example, s1, s2), the information is used and when the information reaches the saturation level TH ( For example, s3, s4, etc.) are interpolated (eg, v3 is obtained) by bit shift or the like using the information of the sample point with the next largest gain (eg, t3 for s3). When it has reached (for example, both s7 and t7 are saturated), interpolation is further performed (for example, v7 is obtained) using information (for example, u7) of the sample point of the minimum gain.

  As described above, by performing A / D conversion and performing interpolation so as to sample by switching a plurality of gains, minute fluctuations detected by the sensor are accurately detected with a very large resolution. The microcomputer detects small fluctuations with medium resolution, and the resolution of very large fluctuations is rough, but the amount of information corresponding to the magnitude of fluctuations is calculated by a microcomputer for camera shake correction. Thus, it is possible to realize both the detection of minute fluctuations with high accuracy and the coverage of a wide fluctuation range. The switching of three gains has been described above, but the switching of four or more gains can be similarly realized.

  Further, in the above-described embodiments, the transmission system having a finite dynamic range such as the A / D conversion units 41 and 42 has been described so as to transmit particularly small amplitude signals with high precision. The present invention is not limited thereto, and can be applied to, for example, a transmission system using a non-linear circuit described below.

  FIG. 13 is a block diagram showing another embodiment of the main part of the present invention. In the figure, the same components as those in FIG. 1A are denoted by the same reference numerals, and description thereof will be omitted. In the embodiment shown in FIG. 13, the sensor signals passing through the coring circuits 71 and 72 and the delay circuit 51 without passing through the coring circuits 71 and 72 are used by using coring circuits 71 and 72 as a kind of non-linear circuit. , 52 is switched by the SW 46 to the sensor signal delayed. The input / output characteristics of the coring circuits 71 and 72 have a dead band for a small input amplitude, for example, as shown in FIG. 14A, and the slope of the input / output characteristics other than the dead band is shown in FIG. It is set to 1.

  FIG. 15 is a signal waveform diagram for explaining the operation of the circuit shown in FIG. The signal waveforms shown in FIGS. 15A to 15F correspond to the signal waveforms shown in FIGS. 5A to 5E. FIG. 15A shows a large-amplitude sensor signal waveform output from the amplifier 39 that does not pass through the coring circuit 71 (or 72), and FIG. 15B shows a coring to which the large-amplitude sensor signal is supplied. 6 shows an output sensor signal waveform of the circuit 71 (or 72).

  These sensor signals are alternately switched by the SW 46 in FIG. 13, output as shown in FIG. 15C, input to the A / D converter 41 (or 42), and are sampled in synchronization with the switching. . Here, assuming that the limit of the dynamic range of the A / D converter 41 is TH, the large amplitude signal is saturated at TH, so that the signal waveform as shown in FIG. You. Therefore, FIG. 15E shows the sampling points.

  Here, similarly to FIG. 5, sampling points q5, q6, q7, q8 at which the large-amplitude signal waveform is saturated at the limit (saturation level) TH of the dynamic range of the A / D converter 41 (or 42). Are interpolated using the output sensor signals of the coring circuit 71 (or 72) at the sampling points n5 to n9.

For example, regarding the signal at the sampling point q5 in FIG.
m5 = TH + {(n5 + n6) / 2}
M5 calculated based on is created and interpolated.

  In the above description, since the slope of the input / output characteristics other than the non-linear dead zone is set to 1 as shown in FIG. 14A, signals up to twice the threshold TH can be passed. If there is a possibility that a signal having an amplitude larger than twice TH is input to the input of the A / D converters 41 and 42, as shown in FIG. If the slope of the output characteristic is made smaller than 1, a larger amplitude can be passed below the TH and passed through the coring circuits 71 and 72. It is possible to cover the passage up to the above large amplitude.

1 is a schematic block diagram of a video camera according to an embodiment of the present invention and a block diagram of main parts. FIG. 2 is a diagram illustrating processing of a camera shake sensor signal of the video camera in FIG. 1. 2 is a time chart for explaining an operation of capturing a camera shake sensor signal of the video camera in FIG. 1. 5 is a flowchart of an embodiment for generating a camera shake correction signal according to the present invention. FIG. 2 is a signal waveform diagram for explaining the operation of FIG. 1 in more detail with respect to a large peak portion in a camera shake sensor signal of the video camera of FIG. 1. 6 is a time chart for explaining an example of an operation of capturing a camera shake sensor signal in FIG. 5. 6 is a time chart for explaining another example of the operation of capturing the camera shake sensor signal in FIG. 5. FIG. 3 is a block diagram illustrating an example of a main part of the video camera of the present invention, in which the time points of an interpolation signal and an interpolated signal are adjusted so as not to cause a time lag. FIG. 11 is a block diagram of another example of adjusting the time points of an interpolation signal and an interpolated signal so as not to cause a time lag in a main part of the video camera of the present invention. FIG. 14 is a block diagram of another example of obtaining a sensor signal having two different gains in a main part of the video camera of the present invention. It is a block diagram of an example of an important section of a video camera of the present invention using a sensor signal of three gains. FIG. 12 is a diagram illustrating an example of a signal waveform and a sampling point of a main part of FIG. 11. FIG. 11 is a block diagram of still another embodiment of a main part of the video camera of the present invention. FIG. 14 is a diagram illustrating each example of input / output characteristics of the coring circuit in FIG. 13. FIG. 14 is a signal waveform diagram illustrating the operation of FIG. 13 in more detail with respect to a large peak portion in a camera shake sensor signal of the video camera of FIG. 13.

Explanation of reference numerals

31 CCD (Solid-state imaging device)
32, 41, 42 A / D converter 33 Signal processing circuit 34, 54, 55 Memory 35 D / A converter 36 Monitor 37, 38 Camera shake sensor 39, 40, 44, 45, 201, 202, 203 Amplifier 43 Microcomputer (Microcomputer)
43a, 43g Sensor signal processing unit 43b Bit extension unit 43c, 46, 206 Switch (SW)
43d Camera shake correction signal generation unit 43e Data control unit 43f Switch control (SWCTL) unit 47-52, 204, 205 Delay circuit 71, 72 Coring circuit 101, 102 AM modulator

Claims (4)

An information signal processing method for transmitting an input analog signal in a transmission system having a finite dynamic range,
A first step of converting the input analog signal into a plurality of analog signals to which different gains have been given, and transmitting the transmission signal through the transmission system;
A second step of separately sampling the plurality of transmitted analog signals at a predetermined sampling clock to generate a plurality of digital signals;
When the values of the sampling points of the plurality of digital signals do not exceed the dynamic range, the values of the sampling points of the digital signals to which the maximum gain has been applied are selected, and the first or two or more digital signal values are selected. When the value of the sampling point exceeds the dynamic range, the digital signal provided with the largest gain among the digital signals not exceeding the dynamic range is adjacent to one or both sides of the first sampling point of the digital signal. Selecting a value interpolated using the value of the second sampling point, and outputting a signal obtained by synthesizing the value of the sampling point selected in the third step in time series. Characteristic information signal processing method. An information signal processing method for transmitting an input analog signal in a transmission system having a finite dynamic range,
Converting the input analog signal into a first analog signal obtained by enlarging a signal portion having a small amplitude equal to or less than a predetermined value, and a second analog signal representing the amplitude of a signal portion having a larger amplitude than the predetermined value; A first step of transmitting the transmission system;
A second step of separately sampling the transmitted first and second analog signals with a predetermined sampling clock to generate first and second digital signals;
When the value of the sampling point of the first digital signal does not exceed the dynamic range, the value of the sampling point is selected, and the value of the first sampling point of the first digital signal falls within the dynamic range. If it exceeds, the dynamic range is larger than the dynamic range calculated by a predetermined arithmetic expression using the values of the second sampling points adjacent to both sides of the first sampling point of the second digital signal. A third step of selecting an interpolated value, and outputting a signal in which the values of the sampling points selected in the third step are synthesized in time series. While temporarily storing the video signal obtained by imaging with the image sensor in the video signal memory, the sensor signal obtained by the angular velocity sensor is transmitted in a transmission system with a finite dynamic range, and the transmitted sensor signal A video camera that reads a video signal subjected to camera shake correction from the video signal memory by variably controlling a reading position of the video signal memory based on the obtained camera shake correction signal,
Conversion means for converting a sensor signal obtained by the angular velocity sensor into a plurality of sensor signals to which different gains are provided,
A / D conversion means for separately sampling the plurality of sensor signals at a predetermined sampling clock to generate a plurality of digital sensor signals,
When the value of the sampling point of the plurality of digital sensor signals does not exceed the dynamic range, the value of the sampling point of the digital sensor signal having the maximum gain is selected, and the first of the one or more digital sensor signals is selected. When the value of the sampling point exceeds the dynamic range, the digital sensor signal having the largest gain among the digital sensor signals not exceeding the dynamic range has one of the digital sensor signals adjacent to one or both sides of the first sampling point of the first sampling point. Selecting means for selecting a value interpolated using the value of the sampling point of 2;
A camera shake correction signal generation unit that generates the camera shake correction signal based on a time-series synthesized signal of the values of the sampling points selected by the selection unit. While temporarily storing the video signal obtained by imaging with the image sensor in the video signal memory, the sensor signal obtained by the angular velocity sensor is transmitted in a transmission system with a finite dynamic range, and the transmitted sensor signal A video camera that reads a video signal subjected to camera shake correction from the video signal memory by variably controlling a reading position of the video signal memory based on the obtained camera shake correction signal,
Conversion means for converting the sensor signal into a first sensor signal obtained by enlarging a signal portion having a small amplitude equal to or smaller than a predetermined value, and a second sensor signal representing the amplitude of a signal portion having a larger amplitude than the predetermined value;
A / D conversion means for separately sampling the first and second sensor signals at a predetermined sampling clock to generate first and second digital sensor signals,
When the value of the sampling point of the first digital sensor signal does not exceed the dynamic range, the value of the sampling point is selected, and the value of the first sampling point of the first digital sensor signal is set to the dynamic range. When the value exceeds the range, the dynamic range calculated by a predetermined arithmetic expression using values of second sampling points adjacent to both sides of the first sampling point of the second digital sensor signal. Selecting means for selecting an interpolation value greater than
A camera shake correction signal generation unit that generates the camera shake correction signal based on a time-series synthesized signal of the values of the sampling points selected by the selection unit.

Images