JP5782546B2 - VIDEO PROCESSING DEVICE, VIDEO PROCESSING METHOD, IMAGING DEVICE, PROGRAM, AND STORAGE MEDIUM - Google Patents

Description

  The present invention relates to a video processing apparatus and a video processing method, and further relates to a method of performing signal processing by detecting a shooting scene from an imaging signal of an imaging apparatus.

  An imaging device such as an electronic camera is provided with an automatic exposure control device that automatically sets the exposure to an optimum value. However, when shooting night scenes and the like, the brightness difference between the subject and the background is large, so an appropriate photometric value cannot be obtained, and an appropriate exposure setting cannot be made. If a night scene is shot with an inappropriate exposure setting, there is a problem that a color skip such as neon of the subject occurs, resulting in an unnatural image. In order to solve this problem, a technique has been proposed in which a histogram of luminance levels is detected and a night view is determined from the shape of the histogram (see Patent Document 1).

  Hereinafter, the night view detection method according to the proposal will be described with reference to FIG. FIG. 2 is an explanatory diagram of a night scene histogram. In this proposed method, a case where a distribution is made only at an extremely high luminance level and an extremely low luminance level as in the solid line histogram shown in FIG. 2 is determined as a night scene. In addition, in order to more accurately determine the night scene state, control for determining the night scene state only when the gain is very high or when the information of the aperture encoder is close to the full aperture is also included. When it is determined that the scene is a night view, exposure control is performed in the direction in which the aperture is reduced, thereby obtaining an optimal captured image with no color jump.

Japanese Patent Application No. 3-263870

  However, the proposed method has the following problems. For example, a histogram when a backlit scene is photographed is likely to be distributed only at an extremely high luminance level and an extremely low luminance level. A similar histogram distribution can be obtained even for a subject with high contrast. On the other hand, in a bright or dark night view, there may be a case where the histogram distribution does not enter the expected threshold level. In such a case, in the proposed method that determines the night scene from the shape of the histogram of the brightness level, the accuracy of determining the night scene deteriorates, and a scene other than the night scene is erroneously determined as the night scene, or even though it is a night scene. There is a problem that the night view cannot be detected.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to make it possible to more accurately determine that a night scene.

In order to achieve the above object, a video processing apparatus according to an aspect of the present invention has the following arrangement. That is,
First acquisition means for acquiring a luminance distribution from the luminance signal of the video imaged by the imaging means;
Second acquisition means for acquiring an evaluation value of contrast based on a high-frequency component of a luminance signal of the video;
A determination unit configured to perform scene determination based on the luminance distribution and the determination result of the point light source based on the evaluation value;

  According to the video processing device and the video processing method of the present invention, it is possible to more accurately determine that the scene is a night view.

The block diagram of the video camera of 1st Embodiment. FIG. 10 is a diagram illustrating an example of a histogram for detecting a night view with the imaging apparatus. The figure which shows the example of a graph of a line peak integral evaluation value. The figure explaining AF evaluation value of a point light source. The figure which shows the example of the histogram in night view photography. The block diagram of the video camera of 2nd Embodiment. The figure which shows the center part which calculates the histogram of an image | video. The flowchart of the night view determination process by embodiment.

<First Embodiment>
FIG. 1 is a block diagram of an imaging system of a video camera according to the first embodiment. In FIG. 1, light incident on a lens 100 is adjusted to an appropriate exposure by an optical diaphragm 101 that adjusts an exposure amount, and is photoelectrically converted by an image sensor 102. The image sensor 102 includes an image sensor such as a CCD or a CMOS. The video signal photoelectrically converted by the image sensor 102 is input to the signal processing device 104 after being converted to an appropriate level by a gain amplifier 103 that changes the gain. The signal processing device 104 performs gamma processing, contour enhancement processing, noise reduction processing, white balance processing, color balance processing, encoding processing according to various recording formats, and the like on the luminance signal and color signal of the input signal. A video signal obtained as a result of signal processing in the signal processing device 104 is recorded as a moving image in the recording device 105. On the other hand, the video signal is also input and displayed on the display device 106 to monitor the captured image at the time of shooting. In addition, the display device 106 displays an OSD display of the state of the imaging device, various settings by the user, and the like over the image being captured.

  Next, a part related to exposure control of the imaging apparatus will be described. In the exposure control of this embodiment, the exposure is evaluated based on the brightness level of the captured image, and feedback control is performed from the amount of incident light to the accumulation time, and the amplifier gain of the imaged signal, etc., to optimize the video signal level Control to keep on. The photometric value acquisition device 107 acquires exposure information from the signal processing device 104 based on a luminance signal in the middle of signal processing. More specifically, a luminance level signal integrated for each of a plurality of regions is acquired as a photometric value from the video signal being processed.

  The evaluation value calculating device 108 calculates an evaluation value for exposure control based on the photometric value acquired by the photometric value acquiring device 107. The exposure evaluation based on the video signal can be performed by, for example, center-weighted average metering with a weight in the center of the imaging area, peak metering using only bright parts in multiple areas, or average metering that averages the entire area. There are various evaluation methods. Among them, it is a common technique to switch the calculation method depending on the shooting mode and the like. The evaluation value calculation device 108 further calculates a target level for exposure control in the exposure control device 109. The target level may change depending on the condition of the subject and the photometric method, and is set as a level at which the evaluation value should be reached. For example, when the evaluation value is lower than the target level, the exposure control device 109 is instructed to increase the exposure amount. Conversely, when the evaluation value is higher than the target level, the exposure control device 109 is instructed to decrease the exposure amount.

  The exposure control device 109 determines exposure control parameters such as an aperture value, shutter speed (charge accumulation time), and amplifier gain with respect to the exposure control direction instructed from the evaluation value calculation device 108. As for exposure control, a program AE method is generally used in which control is performed by a combination of control parameters determined in advance according to the brightness of the subject, but different parameter combinations may be selected depending on subject conditions and photometric methods. Of the exposure control parameters determined by the exposure control device 109, the aperture value is used to control the aperture mechanism driving device 116, the charge accumulation time is used to control the sensor driving device 117, and the amplifier gain is used to control the gain amplifier 103.

  The actuator 115 mechanically operates the optical aperture 101 to change the aperture value. The actuator 115 is driven by the diaphragm mechanism driving device 116, and includes a galvanometer, a stepping motor, or the like. The diaphragm mechanism driving device 116 is a driving device that matches the configuration of the actuator 115, and drives the actuator 115 by current control, pulse control, or the like in accordance with an instruction from the exposure control device 109. The sensor driving device 117 controls the accumulation time of the image sensor 102 and gives the charge accumulation timing, the charge read timing, and the like to the image sensor 102.

  Hereinafter, the configuration and processing relating to the determination of whether or not it is a night view (hereinafter referred to as night view determination) according to the present embodiment will be described. FIG. 8 is a flowchart showing night scene determination processing according to this embodiment. For night scene determination according to the present embodiment, two types of information are used: “Histogram shape of luminance signal” and “Point light source determination”. First, night scene determination (S11 to S14) performed from the histogram shape of the luminance signal will be described. In this night scene determination, whether or not the video is a night scene from the first acquisition process of acquiring a histogram of luminance values from the video captured by the imaging unit including the image sensor 102 and the shape of the histogram acquired in the first acquisition process. This is done by the first determination process for determining whether or not. The histogram generation device 110 that executes the first acquisition processing acquires the luminance signal of the image being processed from the signal processing device 104 (S11). Then, the histogram generation device 110 generates a histogram distribution of luminance levels from luminance signals of pixels in the entire area or almost all areas of the acquired video (S12). The histogram information generated by the histogram generation device 110 is input to the histogram shape determination device 111. The histogram shape determination device 111 determines whether or not night scene shooting is performed from the luminance level distribution pattern from the histogram information. In this embodiment, an object formed with a histogram having peaks in the low-luminance portion and the high-luminance portion is determined to be a night scene (S13, S14). Details of the histogram shape will be described later. The result determined by the histogram shape determination device 111 is finally used for determination by the night view determination device 114.

  Next, a flow (S15 to S18) regarding night view determination based on point light source determination which is the second night view determination of the present embodiment will be described. In the night scene determination based on the point light source determination, the second acquisition process for acquiring the evaluation value of the contrast based on the high frequency component of the luminance signal of the video, and whether or not the point light source subject is included in the video based on the evaluation value. And a second determination process for determining whether the video is a night view. Although not shown in FIG. 1, a so-called TV-AF method is generally used as an autofocus (AF) device used in video equipment such as a video camera. In this method, a high-frequency component is extracted from a video signal obtained from an image sensor, and focus adjustment is performed by driving the focus lens in the optical axis direction so that the high-frequency component becomes large. As an evaluation value of the TV-AF method, there is a line peak integration evaluation value obtained by integrating the peak value of the evaluation value for each horizontal line in the vertical direction. The line peak integration evaluation value is a stable evaluation value with little noise due to the integration effect, is not easily affected by instantaneous noise, and the signal changes sensitively with a slight focus shift. (Determining the moving direction). In the autofocus device, the line peak integral evaluation value in the AF frame is used, but in the following night scene determination, the line peak integral evaluation value for the entire area of the video is used. For example, when an area that does not require discrimination of the point light source is known in advance, the line peak integral evaluation value for the necessary partial area instead of the entire area may be used. Alternatively, the line peak integration evaluation value in a predetermined partial area may be used instead of the entire area of the video.

  The AF evaluation value acquisition device 112 acquires a luminance signal from the signal processing device 104, extracts a high frequency component from the luminance signal, and integrates the peak value of the evaluation value for each horizontal line in the vertical direction (peak value of all horizontal lines). The total of the line peak integral evaluation value is acquired (S15). Although details will be described later, the maximum point of the line peak integral evaluation value is usually the in-focus position, but when shooting a point light source object, the maximum point of the line peak integral evaluation value does not become the in-focus position. is there. The point light source determination unit 113 acquires a change curve of the evaluation value by acquiring the line peak integration evaluation value from the AF evaluation value acquisition device 112 while the autofocus device drives the focus lens in the optical axis direction, Analyze (S16). Then, the point light source determination unit 113 determines whether or not the point light source of the subject being imaged is included based on the evaluation value change curve, and determines that it is a night scene when the point light source is included. In this determination, the above-mentioned characteristic that “the maximum point of the line peak integral evaluation value is not the in-focus position with a point light source” is used. The result determined by the point light source determination unit 113 is used for the final night scene determination in the night scene determination apparatus 114 together with the result determined by the histogram shape determination apparatus 111.

  The night scene determination device 114 determines whether the photographed scene is a night scene based on the determination based on the above-described histogram shape and the point light source determination. More specifically, when both the determination results in the above-described histogram shape determination device 111 and point light source determination unit 113 are night scenes, it is finally determined that the scene is a night scene (S14, S18). When the night scene determination device 114 determines that the night scene is present, the evaluation value calculation device 108 performs evaluation value calculation and target value setting suitable for shooting a night scene (S19). On the other hand, if it is determined that it is not a night scene in at least one of the histogram shape and the night scene determination based on the point light source determination, evaluation value calculation and target value setting other than night scene shooting are performed (S20). This completes the description of the configuration of the video processing apparatus in the present embodiment. Note that the order of processing of night scene determination based on the shape of the histogram (steps S11 to S14) and night scene determination based on the point light source determination (steps S15 to S18) is not limited to this. For example, night scene determination based on point light source determination may be executed first, or both may be executed simultaneously.

<Determination based on histogram shape>
Next, a method for determining a night scene from the histogram shape of the luminance signal will be described. As described with reference to FIG. 2, in a luminance histogram of a general night scene, values are often distributed between extremely high luminance levels and extremely low luminance levels. An extremely high level of brightness corresponds to an illumination part in a night scene subject. On the contrary, an extremely low level of brightness corresponds to a part of the night scene object that is not illumination.

  FIG. 5A is a graph showing an example of a luminance histogram of a night scene. A curve indicated by a solid line represents a luminance histogram curve 601 of the night scene. A dotted line 602 represents a low luminance threshold TH1 for detecting the frequency distribution of the low luminance portion, and a dotted line 603 represents a high luminance threshold TH2 for detecting the frequency distribution of the high luminance portion. Thus, the histogram is divided into three by the low luminance value range TH1 and the high luminance threshold TH2. As a method of detecting the frequency of extremely high luminance levels and extremely low luminance levels, the number of pixels in each luminance range divided by the low luminance threshold TH1 and the high luminance threshold TH2 is counted as a frequency. A comparison is given. First, in order to detect the frequency of an extremely low luminance level, the number of pixels having a luminance level equal to or lower than the low luminance threshold TH1 is counted. Specifically, the area obtained by integrating the number of pixels in the histogram curve 601 indicated by the diagonal lines below the low luminance threshold TH1 is regarded as the area of the non-illuminated part. If this is normalized (averaged) within the luminance level range (TH1-0), the average value of the frequency in the low luminance range is obtained, and the evaluation value of the low luminance level is obtained.

Similarly, in order to detect the frequency of the medium luminance level, the number of pixels of the luminance level in the range larger than the low luminance threshold TH1 and less than the high luminance threshold TH2 is counted. If this is normalized (averaged) within the range of the luminance level (TH2-TH1), an evaluation value of the medium luminance level is obtained. Further, in order to detect the frequency of the high luminance level, the number of pixels of the luminance level in the range of the high luminance threshold TH2 or more is counted. If this is normalized (averaged) within the luminance level range (TH2-100), an evaluation value of a high luminance level is obtained. In order to determine that the histogram-shaped subject as shown in FIG. 5A is distributed only at an extremely high luminance level and an extremely low luminance level,
(Low luminance level evaluation value> Medium luminance level evaluation value) Condition 1
(High luminance level evaluation value> Medium luminance level evaluation value) Condition 2
In this case, it is sufficient that the conditions 1 and 2 are satisfied. That is, when both condition 1 and condition 2 are satisfied, it is determined that the image is a night scene.

  Up to this point, the method is the same as in the proposed method of Patent Document 1, and it may be determined whether or not a night view is taken according to the above conditions. However, there are various actual night scenes. For example, in a night scene close to an evening scene, the frequency of the medium luminance level is assumed to be high because it is not yet dark. In addition, when the illumination is low, the frequency of the high luminance level is low, and a case where the frequency is lower than the evaluation value of the medium luminance level is assumed. FIG. 5B is a graph showing an example of a histogram of a night scene that has less illumination than the night scene shown in FIG. 5A and is bright overall and frequently has a medium luminance level.

  In order to improve night scene detection accuracy, it is necessary to be able to detect the night scene of the histogram shown in FIG. A curve indicated by a solid line in FIG. 5B represents a luminance histogram curve 701 of a night scene. A dotted line 702 represents the low luminance threshold TH1 for detecting the frequency distribution of the low luminance portion, and a dotted line 703 represents the high luminance threshold TH2 for detecting the frequency distribution of the high luminance portion. In FIGS. 5A and 5B, the histogram curve 701 is different from the histogram curve 601, but the low luminance threshold TH1 and the high luminance threshold TH2 are set to the same luminance level.

By using the method described above with reference to FIG. 5A, for the histogram shown in FIG. 5B, the number of pixels in each part delimited by the low luminance threshold TH1 and the high luminance threshold TH2 is counted. A luminance level evaluation value, a medium luminance level evaluation value, and a high luminance level evaluation value can be acquired. In order to determine that the histogram-shaped subject as shown in FIG. 5B is a night scene, it is first necessary to satisfy “condition 1” described above. Here, in this night scene, there is no need to satisfy “Condition 2” because the distribution of the medium luminance level is large and the distribution of the high luminance level is small with little illumination. However, if the high luminance level evaluation value is higher than the low luminance level evaluation value, it will not be a night view.
(Low luminance level evaluation value> High luminance level evaluation value) Condition 3
Need to be. This can be said even when it is distributed only at extremely high brightness levels and extremely low brightness levels.
(Low luminance level evaluation value> Medium luminance level evaluation value) Condition 1
(Low luminance level evaluation value> High luminance level evaluation value) Condition 3
It can be seen that Condition 1 and Condition 3 are common conditions for detecting a subject in a night scene represented by the histograms of FIGS. Therefore, in this embodiment, it is determined whether or not the night view is under this condition.

  In this way, it is possible to determine a night scene by evaluating the histogram shape of the luminance signal. Therefore, in this embodiment, it is assumed that night scene determination is performed according to the above conditions 1 and 3. In addition, it may be determined that the scene is a night view in the case of “condition 1 and condition 3” or “condition 1 and condition 2”. However, as described above, a backlight scene or the like also has a histogram shape similar to that of a night scene, so that the night scene with high accuracy cannot be detected only with the configuration described so far. In the present embodiment, the histogram shape of the luminance signal is used as one of the conditions for night scene detection, and a condition such as a point light source determination described later is added to realize night scene determination with high accuracy.

<Determination by point light source detection>
Next, a method for determining a point light source from the AF evaluation value will be described. As described above, it is possible to detect a point light source in the TV-AF system using the line peak integration evaluation value obtained by integrating the peak value of the evaluation value for each horizontal line in the vertical direction. The line peak integral evaluation value is a stable evaluation value with little noise due to the integration effect, so it is not easily affected by instantaneous noise, and the signal changes sensitively with a slight focus shift, making it ideal for determining the driving direction of the focus lens .

  FIG. 3A is a graph (change curve of the line peak integral evaluation value) showing the relationship between the line peak integral evaluation value and the focus lens position when a normal subject is photographed. As shown in FIG. 3A, in a normal subject, the contrast of the image is improved and the line peak integral evaluation value is increased as it approaches the in-focus position. Therefore, as the AF control, a point where the line peak integral evaluation value is maximized is found while moving the focus lens position, and this is set as the in-focus position.

  On the other hand, the line peak integration evaluation value may not always have a large lens position as the in-focus position. 4A and 4B, when a subject such as a point light source is photographed, the line peak integral evaluation value is set to the left side for each image photographed at the focused focus lens position and the unfocused focus lens position. FIG. When a subject such as a point light source whose length varies due to focus adjustment is photographed, when the point light source is focused as shown in FIG. 4A, the peak value (Y1) of each line is large. The number of lines on which the subject image is applied decreases, and the line peak integration area decreases. On the other hand, as shown in FIG. 4B, the peak value (Y2, Y3) of each line becomes smaller (Y1> Y2> Y3) when it becomes out of focus with respect to the point light source, but the number of lines on which the subject image is applied. As the value increases, the line peak integral evaluation value increases. That is, the line peak integral evaluation value becomes maximum (Y2 × 2 + Y3 × 2) in the vicinity of the in-focus, and a phenomenon (Y1 × 2 <Y2 × 2 + Y3 × 2) occurs in the in-focus position.

  FIG. 3B is a graph showing the relationship between the line peak integral evaluation value and the focus lens position when a subject such as a point light source is photographed. As shown in FIG. 3B, in a subject such as a point light source, the maximum point of the line peak integral evaluation value is not the in-focus position. For this reason, as AF control, a subject whose line peak integral evaluation value becomes smaller at the in-focus position, such as a point light source, is discriminated, and automatic focusing control can be switched to ensure in-focus.

  The point light source determination unit 113 of the present embodiment uses the above-mentioned characteristic that “the maximum point of the line peak integral evaluation value does not become the in-focus position with a point light source”, and performs imaging based on the line peak integral evaluation value. It is determined whether or not the point light source of the subject is included.

  Further, the line peak integral evaluation value is used as the AF evaluation value. However, in the case of a point light source subject as shown in FIG. 4, the characteristic line peak integral evaluation value as shown in FIG. Distribution. Using this, it is possible to determine a point light source also from the relationship between a high luminance line and a low contrast line. That is, as a feature of a night scene, there are high luminance and high contrast lines composed of point light sources while there are many low luminance and low contrast lines. Therefore, it can be determined that the light source is a point light source from the ratio of the high luminance line and the low contrast line. That is, when the ratio between the number of horizontal lines with the evaluation value smaller than the first threshold and the number of horizontal lines with the evaluation value larger than the first threshold and the second horizontal threshold is less than a predetermined value, the video is displayed. Is determined to include a point light source.

  Furthermore, the determination that the point light source is based on the ratio of the high-contrast line and the low-contrast line described above and the determination using the fact that the line peak integral evaluation value becomes maximum near the in-focus position and falls at the in-focus position may be used in combination. Good. By doing in this way, the determination accuracy of a point light source can further be improved.

  Point light source determination using the line peak integral evaluation value is used as one of the conditions for night scene detection. By using it together with the above-described night scene detection condition based on the histogram shape of the luminance signal, night scene determination with high accuracy is possible. Can be realized. In the above description, the line peak integration evaluation value over the entire area of the video is used. However, the line peak integration evaluation value for the video in the AF frame used by the autofocus device may be used. In that case, the line peak integral evaluation value acquired by the autofocus device for focus position detection can be used as it is. Furthermore, when there are a plurality of AF frames, it may be determined that the scene is a night scene when the characteristics indicating the point light source are detected in at least one of them.

  Next, exposure control when it is determined as a night scene will be described. In general, when a night scene is photographed in normal exposure control that is not conscious of the night scene, there are many dark portions, so that the brightness is controlled to be brighter than the actual subject. As a result, it is necessary to increase the gain, the shot noise increases, and the dark portion becomes brighter and the image is floated. Furthermore, the illumination portion is saturated on the image sensor or the signal processing, and becomes a white flying state.

  Various exposure controls suitable for night scenes can be considered. For example, it is possible to make the photometry method suitable for night scenes. In order to prevent the illumination portion from being blown out, photometry (peak photometry) for generating a photometric evaluation value from the peak level of luminance in the image is effective. By controlling the brightness peak level to a level that does not cause whiteout, it is possible not only to express the color and tone of the illumination, but also to capture dark areas darkly. However, if peak metering is applied to a normal subject, the image becomes darker than the appropriate level. For this reason, when the above-described night scene determination is performed, peak exposure metering is performed, so that an optimum exposure can be automatically obtained from a normal scene to a night scene without any special switching.

  Another example of exposure control suitable for a night scene is to lower the target level of exposure. As described above, when a night scene is photographed in the normal exposure control without being aware of the night scene, there are many dark portions, so that the brightness is controlled to be brighter than the actual subject. In order to avoid this, it is only necessary to lower the target level of exposure, and it is possible to avoid making the dark part of the subject brighter than necessary. A similar effect can be obtained by a technique that suppresses the gain increase by suppressing the exposure control amount, and can be further improved in combination with the above-mentioned peak photometry. If these methods are also applied to a normal subject that is not a night view, the image becomes darker than the appropriate level. For this reason, if the exposure control is automatically switched when the above-mentioned night scene determination is made, the optimum exposure can be automatically obtained from the normal scene to the night scene without any special switching.

  The above is the description regarding night view determination according to the first embodiment. According to the present embodiment, when it is determined that the histogram shape has a mountain in each of the low-luminance portion and the high-luminance portion and there is a point light source in the subject, it is determined to be a night scene. For this reason, it is possible to correctly determine whether or not the image is a night scene even if the image is erroneously determined by imaging control based on general night scene determination (for example, when a backlight scene or the like is captured). . For this reason, it is possible to appropriately switch to exposure control specialized for night views, and to obtain an optimal image. Furthermore, it is possible to detect a night scene even if the night scene has few features such as a bright night scene.

  In the present embodiment, the line peak integration evaluation value for AF control has been described. However, the point light source may be configured exclusively for night scene detection, or the point light source may be detected in units of pixels instead of lines. Easily possible. Further, in the present embodiment, the configuration in which the exposure control is changed according to the night scene determination has been described. However, changing the luminance γ curve, the setup level, and the color signal processing may also achieve the effect of night scene shooting. Needless to say, it is good.

Second Embodiment
In the first embodiment, night scene determination is performed using two types of information: “histogram shape of luminance signal” and “point light source determination”. In the second embodiment, instead of “point light source determination”, “ratio determination of high-luminance pixels in the center (described later)” is used. FIG. 6 is a block diagram of an imaging system of a video camera according to the second embodiment. In FIG. 6, the same reference numerals are attached to the same components as those of the video camera (FIG. 1) of the first embodiment. The gate unit 812 extracts only the frequency at the center of the image from the histogram generated by the histogram generation device 110. The high luminance part counting device 813 calculates the ratio of the number of pixels of the high luminance part in the central part of the image extracted by the gate part 812 to the entire image. The night scene determination device 814 determines whether the photographed scene is a night scene based on the night scene determination by the histogram shape determination device and the night scene determination based on the ratio of the high brightness pixels in the center. Note that exposure control is the same as in the first embodiment.

  The configuration related to night view determination according to the second embodiment will be described below. Night scene determination according to the second embodiment is made from two types of information, one of which is determined from the histogram shape of the luminance signal (first acquisition processing and second determination processing), as in the first embodiment. It is. Therefore, since the flow of generating a histogram and determining the night view has been described in the first embodiment (steps S11 to S14), it is omitted here. The result determined by the histogram shape determination device 111 is finally used for determination by the night view determination device 814.

  Next, a flow related to the ratio determination of the high-intensity pixels at the center, which is the second night scene determination in the second embodiment, will be described. In the second embodiment, the second acquisition process and the second determination process for performing night scene determination using the point light source of the first embodiment are replaced with the determination based on the ratio of the high-intensity pixels in the center. That is, in the second acquisition process, the number of pixels exceeding a predetermined luminance value is acquired inside the partial area set at the center of the video, and in the second determination process, all the videos of the acquired number of pixels are acquired. It is determined whether or not the video is a night scene based on the ratio to the number of pixels. Note that this night scene determination processing is executed in steps S15 to S18 of FIG.

  There are various images of night scenes, but if there is no illumination at the center of the image, there is a high possibility that another subject will be present there. Alternatively, such a subject may actually be a backlight scene instead of a night view, and there is a high possibility that the photographer intends to shoot with the main subject in the center. When exposure control suitable for night scenes is performed on such a subject, the central portion is crushed in black, resulting in an image with no gradation for the main subject. Based on the above, if “the presence of a high-intensity part that is equal to or higher than the threshold value is present in the central part of the video” is added as a condition for the night scene, false detection is reduced and detection accuracy is improved.

  The gate unit 812 extracts information on only the central portion of the video from the histogram of the luminance signal level generated by the histogram generation device 110 (second acquisition process). Then, the gate unit 812 generates a histogram specialized for the central portion as preparation for detecting whether or not there is an illumination in the central portion in the video. The high brightness portion counting device 813 counts the number of pixels in the high brightness portion in the histogram specialized for the central portion generated by the gate portion 812. More specifically, the high luminance part counting device 813 counts the number of pixels in the histogram that exceeds a predetermined luminance level threshold value, and calculates the number of high luminance pixels in the central portion relative to the total number of pixels in the video. Calculate the percentage. The high luminance part counting device 813 further compares the calculated ratio of the number of high luminance pixels with a threshold value, and determines that the ratio is equal to or greater than the threshold value is a night scene (second determination process). The result determined by the high luminance part counting device 813 is used together with the result determined by the histogram shape determining device 111 for the final determination by the night scene determining device 814.

  The night scene determination device 814 finally determines that it is a night scene when both the result determined by the above-described histogram shape determination device 111 and the result determined by the high luminance part counting device 813 are night scenes. When the night scene determination device 814 determines that the scene is a night scene, the evaluation value calculation device 108 performs evaluation value calculation and target value setting suitable for night scene photography.

  FIG. 7 is an image diagram showing a central portion of the entire video area handled by the gate unit 812. In the figure, reference numeral 901 denotes an entire video area, and reference numeral 902 denotes a central area. The size and position of the central region 902 are, for example, the horizontal width is about 1/3 of the entire image region 901 and the central position, and the vertical width is about 1/3 of the region 901 and larger than the center. The layout is slightly below. Here, the reason why the arrangement is slightly lower than the center in the vertical direction is because there is a tendency that the subject is generally captured slightly below the screen when shooting a night scene. Of course, the setting of the central portion is not limited to this, and it may be shifted in the horizontal direction. That is, the center of the image and the center of the central area do not have to coincide with each other, and a partial area having a predetermined size and position may be used. This will be described as a region. The gate unit 812 performs processing for extracting only the frequency of the luminance signal located inside the central region 902.

  The high brightness level counted by the high brightness section counting device 813 will be described with reference to FIG. However, the histogram generated by the gate unit 812 is a histogram created for the pixels in the center region 902. The high brightness portion counting device 813 counts the number of pixels having a brightness level equal to or higher than the threshold value TH2. The threshold value TH2 may be the same as or different from the threshold value TH2 used by the histogram shape determination apparatus 111 described above. The high-luminance part counting device 813 determines that a night scene is obtained when the ratio of the number of pixels at the high-luminance level in the central portion to all the pixels exceeds a predetermined threshold A. As the threshold A, a result of investigating various night view scene data shows that about 0.7% is appropriate.

  Night scene determination using the ratio of the number of high-luminance pixels in the central part is used as one of the conditions for night scene detection. A night scene with high accuracy can be obtained by using it together with the above-described night scene detection based on the histogram shape of the luminance signal. Judgment can be realized. Further, by adding the point light source determination using the line peak integration evaluation value described in the first embodiment to the condition, the night scene detection system can be further improved.

  Note that the exposure control when it is determined to be a night scene is the same as in the first embodiment. For example, when the above-described night scene scene determination is performed, peak exposure metering is performed, so that an optimum exposure can be automatically obtained from a normal scene to a night scene scene without performing a special switching operation. Further, as another example of exposure control suitable for a night scene, as described in the first embodiment, the same effect can be obtained by a method of avoiding an increase in gain by reducing the target level of exposure or by controlling the exposure control amount. Can be obtained. In addition, it is possible to further improve the effect in combination with the above-mentioned peak photometry. When these methods are also applied to a normal subject other than the night view, the video becomes darker than the appropriate level. For this reason, if the exposure control is automatically switched when the above-mentioned night scene determination is made, the optimum exposure can be automatically obtained from the normal scene to the night scene without any special switching.

  The above is the description regarding night view determination according to the second embodiment. According to the second embodiment, the same effect as the first embodiment can be obtained.

  In the second embodiment, the gate unit 812 has been described with the configuration in which the central portion is extracted after generating the histogram. However, the central portion of the video is extracted to generate the luminance histogram of the central portion. Also good. In addition, the gate unit 812 does not extract only the central portion, but the luminance value obtained by taking the central portion weighted average of the luminance values in which the central portion is weighted and the peripheral portion is taken into consideration exceeds a predetermined threshold value. Whether or not the night view is determined may be determined. That is, the average luminance value of the luminance value in the video is calculated by multiplying the luminance value of the pixel inside the partial area set at the center of the video by a weight value larger than that outside the partial area. If it is larger than the low value, it may be determined that it is a night view. The weight value is preferably, for example, the weight of the central region: the weight outside the central region = 2: 1, but is not limited thereto.

  Further, as shown in the third embodiment, night scene determination based on detection of a point light source described in the first embodiment may be added. That is, the acquisition of the line peak integral evaluation value described in the first embodiment and the night scene determination based on the detection of the point light source based on the evaluation value are added as the third acquisition process and the third determination process. If all of these are determined to be night views, they may be determined to be night views.

<Third Embodiment>
According to the first and second embodiments described above, the point light source determination using the line peak integral evaluation value, the histogram shape of the luminance signal, and the ratio of the number of high luminance pixels in the center are combined. It is possible to realize night scene determination with high accuracy. However, even if the night scene is detected with high accuracy, if it is detected with an evaluation value close to the threshold value of each condition, it may be determined that the scene is not a night scene due to a slight change in the subject. Considering exposure control when it is determined that the scene is a night view, frequent changes in the detected scene (switching whether or not the scene is a night scene) results in an unstable exposure level, which is not desirable for an imaging apparatus. Therefore, in the third embodiment, hysteresis is provided for each threshold, and the above problem is solved by making the night scene determination condition different from the determination condition for exiting the state determined to be the night scene.

In the third embodiment, a determination condition for determining that it is not a night view after it is determined once is a night view will be described. Assuming that the condition for determining a night view is an “IN condition” and the condition for determining that the night view is no longer a night scene is an “OUT condition”, the above-described night scene determination conditions are defined as follows.
Judgment by point light source: Judgment A
Judgment by histogram shape: Judgment B
Judgment based on the ratio of the number of high-luminance pixels in the center: Judgment C
In this case, in order to increase the detection accuracy, the logical product of the respective determination conditions is set as the IN condition. That is,
IN condition: (Judgment A and Judgment B and Judgment C)
As an IN condition, it is desirable to determine a night scene when any of the determination conditions is a night scene (affirmative).

However, if any one of the determination conditions is missing from the state in which the night view is determined under the above IN condition (if any determination result is negative), the determination is changed. Becomes frequent. Therefore, once a night view is determined under these conditions, stable control can be achieved by changing the determination that it is not a night view only when all conditions cannot be satisfied as the OUT condition. That means
OUT condition: (not judgment A and not judgment B and not judgment C)
Is an OUT condition, and the determination that the scene is a night scene is maintained until this condition is satisfied. However, this is an example. For example,
OUT condition: (not judgment A and (not judgment B or judgment C))
And so on. In other words, the conditions may be looser than the IN conditions, and the combinations vary. That is, the feature of this embodiment is that, in night scene detection having a plurality of determination conditions, the combination of determinations under conditions (OUT conditions) that deviate from the night scene detection state is different from the combination of determinations under IN conditions. is there.

  The above is the description regarding redetermination after night scene detection according to the present invention. According to the present embodiment, in conventional imaging control based on night scene detection, even when night scene detection becomes unstable due to subject change or panning, it can be stably determined that the scene is a night scene. .

  In the present embodiment, the above-described three types of determination means have been described, but the number of determination means is not limited as long as it is plural. Further, a condition that is not in the IN condition may be added to the OUT condition, and conversely, the condition that is in the IN condition may not be in the OUT condition. When two determination conditions are used as in the first and second embodiments, the determination that the scene is a night scene may be maintained until the result of both determinations is negative.

As described above, according to the present invention, when a backlight scene is photographed or when a subject with high contrast is photographed, it is prevented from being erroneously determined to be a night scene, and the night scene determination accuracy is improved. Can be made. Furthermore, it is possible to automatically obtain a captured image having an optimum luminance level without color jump when a night scene is photographed.
In the above embodiment, application to a digital video camera has been described. However, it can be used in the same manner as a digital still camera. In this case, the above-described processing can be performed at the time of autofocus processing in the digital camera, and the night scene shooting mode can be automatically set. Further, the present invention is not limited to an imaging apparatus, and the present invention can also be applied to a reproduction apparatus such as a digital television that reproduces a moving image file and application software executed on a personal computer. In this case, for each predetermined frame, the luminance value histogram of the video luminance signal and the evaluation value of the contrast are recorded in a moving image file and read out and used, or the reproduced video signal of the moving image is analyzed. There is a method for acquiring such information directly.

Although the present invention has been described in detail based on preferred embodiments thereof, the present invention is not limited to these specific embodiments, and various forms within the scope of the present invention are also included in the present invention. included. A part of the above-described embodiments may be appropriately combined.
The control of the system control circuit 1 may be performed by one hardware, or the entire apparatus may be controlled by a plurality of hardware sharing the processing.
Although the present invention has been described in detail based on the preferred embodiments thereof, the present invention is not limited to these specific embodiments, and various forms without departing from the gist of the present invention are also included in the present invention. included. Furthermore, each embodiment mentioned above shows only one embodiment of this invention, and it is also possible to combine each embodiment suitably.
Further, in the above-described embodiment, the case where the present invention is applied to a digital camera has been described as an example, but this is not limited to this example. That is, the present invention displays a plurality of images at the same time, such as a personal computer, a PDA, a mobile phone terminal, a portable image viewer, a display for selecting and confirming a print image provided in a printer, and a digital photo frame. Any display control device that can be controlled is applicable.

(Other embodiments)
The present invention is also realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program code. It is a process to be executed. In this case, the program and the storage medium storing the program constitute the present invention.

Claims (13)

First acquisition means for acquiring a luminance distribution from the luminance signal of the video imaged by the imaging means;
Second acquisition means for acquiring an evaluation value of contrast based on a high-frequency component of a luminance signal of the video;
An image processing apparatus comprising: a determination unit configured to perform scene determination based on the luminance distribution and a determination result of a point light source based on the evaluation value. The luminance distribution acquired by the first acquisition unit is a histogram,
The determination means includes first determination means for determining whether or not the shape of the histogram acquired by the first acquisition means satisfies a predetermined condition;
Second determination means for determining whether or not a condition including a subject of a point light source is included in the video based on the evaluation value acquired by the second acquisition means,
The video processing apparatus according to claim 1, wherein when the first determination unit and the second determination unit determine that both of the conditions are satisfied, the video processing apparatus determines that the scene is a specific scene.   The video processing apparatus according to claim 2, wherein the specific scene is a night scene. The second acquisition means acquires the peak value of the high-frequency component for each horizontal line of the video, acquires the sum of the peak values for all horizontal lines of the video as the evaluation value,
The second determination unit sandwiches the in-focus position detected by the autofocus unit in a change curve indicating a change in the evaluation value accompanying the movement of the focus lens in the optical axis direction by the autofocus unit included in the photographing unit. 4. The video processing apparatus according to claim 2, wherein when the two convex peaks exist, it is determined that the video includes a point light source. 5. The second acquisition means acquires the peak value of the high frequency component for each horizontal line of the video as the evaluation value,
Said second judging means, wherein the evaluation value based on the number of first small and the number of horizontal lines than the threshold value, greater horizontal lines than the evaluation value before Symbol first second threshold greater than the threshold value 4. The video processing apparatus according to claim 2, wherein it is determined whether or not a point light source is included in the video.   The second determination means is configured such that a ratio of the number of horizontal lines whose evaluation value is larger than a second threshold value to the number of horizontal lines whose evaluation value is smaller than a first threshold value is predetermined. The video processing apparatus according to claim 5, wherein it is determined that a point light source is included in the video when the value is equal to or less than the value.   After the determination means determines that the video is the specific scene, the first determination means determines that the shape of the histogram does not satisfy a predetermined condition, and the second determination means determines the subject of the light source 7. The video processing apparatus according to claim 2, wherein the video processing apparatus determines that the scene is the specific scene until it is determined that the condition including the video is not satisfied. 7.   The video processing apparatus according to claim 1, further comprising a calculation unit that calculates a photometric evaluation value corresponding to the scene determined by the determination unit. A focus detection unit that detects a focus position using a contrast evaluation value based on a high-frequency component of a luminance signal of a video imaged by the imaging unit;
The video processing apparatus according to claim 1, wherein the evaluation value acquired by the second acquisition unit is an evaluation value used by the focus detection unit to detect a focus position.   Photographing means;
  The video processing apparatus according to any one of claims 1 to 9,
  An image pickup apparatus comprising: an exposure control unit that controls an exposure of the photographing unit according to a scene determined by the determination unit. A first acquisition step of acquiring a luminance distribution from the luminance signal of the video imaged by the imaging means;
A second acquisition step of acquiring a contrast evaluation value based on a high-frequency component of the luminance signal of the video;
A video processing method comprising: a determination step of performing scene determination based on the luminance distribution and a determination result of a point light source based on the evaluation value. A program for causing a computer to execute the video processing method according to claim 11 .   A computer-readable recording medium on which a program for causing a computer to execute the video processing method according to claim 11 is recorded.