WO2006137216A1 - Imaging device and gradation converting method for imaging method - Google Patents

Abstract

Gradation converting means (8) has a nonlinear conversion characteristic such that when the input is in a first range, the variation of the output with the variation of the input is small, and when the input is in a second range, the variation of the output with the variation of the input is large. Correction control means (7) performs an offset correction of a signal (Sd) in proportion to the imaging output of a solid-state imaging element (1) when the number of components in the first range of the imaging signal (Sf) before the gradation conversion is large and determines the correction value in such a way that the number of components in the second range increases. With this, the difference in brightness between small areas of the image can be clearly reproduced.

Description

 Specification

 Imaging device and P-tone conversion method in imaging device

 Technical field

 TECHNICAL FIELD [0001] The present invention relates to an image pickup apparatus that performs a predetermined gradation conversion process on image data having a signal value for each pixel obtained by an image pickup device, and more particularly to a character on a document, a human face. The present invention relates to an image pickup apparatus and an image pickup method having gradation correction means suitable for detecting small changes in luminance in details of a subject, such as features, blood vessel patterns, and fingerprint patterns.

 Background art

 [0002] As a gradation (correction) control of a conventional imaging device, a plurality of nonlinear gradation conversion characteristics are provided, and the gradation conversion characteristics are selected according to the luminance of the image (maximum brightness average value). Some have realized an expansion of the dynamic range (see, for example, Patent Document 1).

 Patent Document 1: Japanese Patent Application Laid-Open No. 2004-363726 (page 9, FIG. 5)

 Disclosure of the invention

 Problems to be solved by the invention

[0004] However, a small change in the brightness of the details of the image obtained by imaging (detected as a high-frequency image signal) causes a large change in the brightness of the entire screen (as a low-frequency image signal). The dynamic range of the imaging device is optimized for low-frequency signal components when it is superimposed on the image of the subject (especially, such as human facial features, blood vessel patterns, and fingerprint patterns). There was a problem that small brightness differences in image details could not be reproduced clearly.

 Means for solving the problem

[0005] An imaging apparatus according to the present invention includes a solid-state imaging device, a luminance distribution detecting unit that detects a luminance distribution and a maximum level and a minimum level of an imaging signal proportional to an imaging output obtained from the solid-state imaging device. And correction means for performing at least one of offset correction and gain correction on the imaging signal, and the correction based on the maximum level and the minimum level of the luminance distribution detected by the luminance distribution detection means! Correction amount determining means for controlling the correction amount so as to expand the range of change of the corrected imaging signal output from the means. [0006] Further, there is provided a subject shape recognition means for recognizing the shape of the subject on the screen from the feature quantity of the imaging signal proportional to the imaging output obtained from the solid-state imaging device and outputting a subject area signal indicating the subject area. In addition,

 The luminance distribution detecting means detects the luminance distribution in the subject area based on the subject area signal output from the subject shape recognizing means, and the correction amount determining means is the subject area detected by the luminance distribution detecting means. On the basis of the maximum level and the minimum level of the luminance distribution, the correction amount is controlled so as to expand the change range of the corrected imaging signal in the subject area output from the correction means.

 The invention's effect

 [0007] According to the present invention, the offset correction, gain correction, and gradation conversion characteristics of the imaging signal are adaptively controlled based on the luminance distribution of the imaging signal and the maximum and minimum levels. Therefore, a small luminance difference in the details of the image can be reproduced clearly.

 Also, it detects the shape of the subject in the captured image and adapts the offset correction, gain correction, and gradation conversion characteristics of the imaged signal based on the luminance distribution of the imaged signal in the subject area and its maximum and minimum levels. Therefore, the small brightness difference in the details of the subject image can be reproduced more vividly.

 Brief Description of Drawings

FIG. 1 is a block diagram showing an imaging apparatus according to Embodiment 1 of the present invention.

 FIG. 2 (A) to (C) are diagrams showing an example of conversion characteristics of the gradation conversion means of the first embodiment.

 FIG. 3 is a block diagram showing an example of luminance distribution detection means used in the first embodiment.

 4 is a diagram showing an example of a histogram generated by the histogram generation unit of FIG.

 FIG. 5 is a diagram for explaining a method for determining an offset correction amount and a gain correction amount by the correction amount determining means in FIG. 1.

 FIG. 6 is an example of a table used when the gain correction amount is obtained from the minimum level value detected by the luminance distribution detecting means in FIG. 1.

FIG. 7 (A) and (B) are diagrams showing an example of control of the gradation converting means by the correction amount determining means in the first embodiment. 8] (A) and (B) are diagrams showing an example of control of the gradation converting means by the correction amount determining means in the first embodiment.

 [FIG. 9] (A) to (G) are timing charts showing the operation of the imaging apparatus of the first embodiment.

FIG. 10 is a flowchart showing a processing procedure in the first embodiment.

 [FIG. 11] (A) to (C) are diagrams showing the influence of correction according to Embodiment 1 on a signal after gradation conversion.

 FIGS. 12A to 12C are diagrams showing an example of the gradation correction operation according to the first embodiment.

FIG. 13 is a diagram showing an example of screen division in the second embodiment of the present invention.

FIG. 14 is a flowchart showing a processing procedure in the second embodiment.

 FIG. 15 (A) to (G) are timing charts showing the operation of the third embodiment of the present invention.

FIG. 16 is a block diagram showing a configuration of an imaging apparatus according to Embodiment 4 of the present invention.

 FIG. 17 is a flowchart showing a processing procedure in the fourth embodiment.

 [FIG. 18] (A) to (G) are timing charts showing the operation of the imaging apparatus of the fourth embodiment.

 FIG. 19 is a block diagram showing Embodiment 5 of the present invention.

 FIGS. 20A to 20G are timing charts showing the operation of the fifth embodiment.

 FIG. 21 is a flowchart showing a processing procedure in the fifth embodiment.

 FIG. 22 is a block diagram showing an example of luminance distribution means used in the sixth embodiment.

 FIG. 23 (A) and (B) are waveform diagrams showing the operation of the luminance distribution means 31 of FIG.

 24 is a block diagram showing an example of amplitude detection means in FIG.

 FIG. 25 is a diagram showing an example of a screen including areas with different brightness.

 FIGS. 26A to 26C are diagrams showing an example of the gradation correction operation according to the sixth embodiment. Explanation of symbols

1 solid-state imaging device, 2 amplification means, 3 AZD conversion means, 4 exposure control means, 5 imaging signal memory means, 6 correction means, 61 offset correction means, 62 gain correction means, 7 correction control means, 8 gradation conversion means, 9 Image processing means, 10 subjects Shape recognition means, 11 luminance distribution detection means, 12 correction amount determination means.

 BEST MODE FOR CARRYING OUT THE INVENTION

 [0010] Embodiment 1.

 FIG. 1 is a block diagram showing a configuration of an imaging apparatus according to Embodiment 1 of the present invention. The imaging apparatus according to Embodiment 1 includes a solid-state imaging device 1, an amplification unit 2, an AZD conversion unit 3, an exposure control unit 4, an imaging signal memory unit 5, a correction unit 6, and a correction control unit 7. Gradation converting means 8, image processing means 9, and subject shape recognizing means 10.

 The correction unit 6 includes an offset correction unit 61 and a gain correction unit 62, and the correction control unit 7 includes a luminance distribution detection unit 11 and a correction amount determination unit 12.

 [0011] The solid-state imaging device 1 photoelectrically converts light, which has also entered subject power, for each pixel, and outputs an output signal of the solid-state imaging device (sometimes referred to as "imaging output"). The solid-state imaging device 1 has a photoelectric conversion element that forms a plurality of pixels arranged two-dimensionally, that is, in the horizontal direction and the vertical direction, and the photoelectric conversion element has a size corresponding to the amount of incident light. The analog signal is output. The output of the signal from the solid-state imaging device 1 is sequentially output with the pixel forces aligned in the horizontal direction on each horizontal line, and such a signal for each pixel is sometimes called a pixel signal.

 The solid-state imaging device 1 is, for example, a CCD imaging device, and performs control corresponding to shutter speed and shutter opening / closing by controlling the photocharge accumulation time and timing of taking out the accumulated charge. Can do. Therefore, the control of the charge accumulation time for the solid-state imaging device 1 is sometimes called an electronic shutter function. In this embodiment, the exposure control means 4 controls the charge accumulation time.

 [0013] The solid-state imaging device 1 may be a monochrome sensor or a color sensor. The color sensor may be a primary color filter array or a complementary color filter array. Further, the solid-state imaging device 1 may use a CMOS imaging device.

 [0014] The luminance signal Y of the image sensor using the primary color filter is

 Y = 0. 299 XR + 0. 587 X G + 0. 114 X B

 It is obtained by.

In addition, the luminance signal of the image sensor using the complementary color filter is Y = (Cy + Mg) + (Ye + G)

 Or

 Y = (Cy + G) + (Ye + Mg)

 It is obtained by.

 In a monochrome image sensor, the Y signal is used as it is.

The amplification means 2 amplifies the imaging output of the solid-state imaging device 1 and outputs an imaging signal proportional to the imaging output. The output of the amplifying means 2 may be referred to as an amplified output. The amplification gain of the amplification means 2 is controlled by the exposure control means 4.

[0016] The AZD conversion means 3 performs AZD (analog-digital) conversion on the output of the amplification means 2 and outputs a digital signal (sometimes called "AZD output" or "imaging data") Sc. The digital signal represents the luminance for each pixel, and the digital signal for each pixel is sometimes called pixel data. The AZD output Sc is also an imaging signal proportional to the imaging output of the solid-state imaging device 1.

 [0017] The exposure control means 4 determines the exposure amount of the solid-state imaging device 1 and the amplification gain of the amplification means 2 according to the luminance (Y) signal of the AZD output proportional to the imaging output or green (the value of the GH symbol). Control.

 [0018] The exposure amount control of the present embodiment is realized by controlling the charge accumulation time of the solid-state imaging device 1, but instead, the control of the light amount of the light source that illuminates the subject and the opening value of the aperture are adjusted. It may be realized by iris control or a combination of these.

 [0019] The imaging signal memory means 5 has a memory capacity for holding AZD output (imaging data) in units of frames, and was written one frame before writing one frame of imaging signals to the memory. An imaging signal can be read out. The signal Sd read from the imaging signal memory means 5 may be referred to as “imaging signal memory output”.

 [0020] The offset correction means 61 is composed of, for example, a subtracter, and subtracts the offset correction amount (Kb) from the correction amount determination means 12 from the output of the imaging signal memory 5, and the gain correction means Output to 62. The output of the offset correction means 61 may be referred to as “offset correction output”.

The gain correction means 62 is constituted by a multiplier, for example, and the gain correction from the correction amount determination means 12 is performed. Multiply the positive amount (Ka) by the output of the offset correction means 61, and the multiplication result is the gradation conversion means.

Output to 8. The output Sf of the gain correction means 62 may be referred to as a “corrected imaging signal”.

The gradation converting means 8 performs gradation conversion on the corrected imaging signal Sf and outputs it. The output Sg of the gradation conversion means 8 is sometimes called “gradation conversion output”.

 FIG. 2A shows an example of the gradation conversion characteristics of the gradation converting means 8. The horizontal axis is input (referred to as “gradation conversion input”), and the vertical axis is output (referred to as “gradation conversion output”). In the example shown

The input is a 10-bit number, the minimum is 0, the maximum is 1023, and the output is an 8-bit number, the minimum is 0, and the maximum is 255.

[0023] Fig. 2 (B) below the horizontal axis of the gradation conversion characteristics in Fig. 2 (A) shows an example of the input signal Sf, and the output signal Sg for such an input signal is represented by a gradation conversion characteristic curve. It is shown in Fig. 2 (C) on the right side of.

 In the gradation conversion characteristics shown in the figure, in the portion where the input is small (or range), the increase in output relative to the increase in input is relatively large (the slope of the characteristic curve is relatively steep), and the input is large. As a result, the slope of the characteristic curve gradually becomes gentler. In the example shown in the figure, the change of the output with respect to the input is linear when the input is less than the maximum value of approximately 1Z2. This characteristic is advantageous in that a relatively low-brightness part can be reproduced with relatively bright and high contrast, and a relatively high-brightness part can be reproduced with gradation, resulting in an image with a wide dynamic range. There is.

 [0025] It should be noted that the tone conversion characteristics can be γ characteristics, polygonal line characteristics, and the like, and are determined in consideration of various imaging conditions such as monitor characteristics, subject conditions, illumination light conditions, and the like. The present invention emphasizes the change in the signal by making the slope of the characteristic curve steeper than that of the other parts of the input range, that is, the first part where the slope of the characteristic polarity is relatively gentle. This can be applied to the case where it has (range) and the second part (range) where the slope of the characteristic curve is relatively steep.

 The gradation conversion means 8 can be realized by a look-up table (LUT), for example. Further, when the conversion characteristic curve is represented by a broken line, it can be realized by a relatively simple calculation, and can also be realized by hardware logic.

[0027] The output Kc of the correction amount determination means 8 is used to switch the gradation conversion characteristics of the gradation conversion means 8. Switching signal. For example, in the case of a hardware configuration in which the gradation conversion means 8 can hold a plurality of gradation conversion characteristics, it is possible to switch the gradation conversion characteristics in units of frames at high speed. When the gradation conversion characteristic of the gradation converting means 8 is configured by a polygonal line, the output Kc of the correction amount determining means 8 is the coordinate value (input value and output) of the broken point position for realizing the polygonal line characteristic. Value) etc. may be supplied to the gradation converting means 8 in units of frames. In this case, there is an effect that the gate scale and the memory capacity can be reduced because the gradation converting means 8 does not need to have a storage area for realizing a plurality of gradation conversion characteristics.

 [0028] The image processing means 9 includes white balance control, output power of the solid-state image pickup device 1 having a Bayer array or a No-cam array, interpolation signal processing for generating an RGB signal of each pixel, noise removal, RGB signal It has functions to realize general camera signal processing, such as conversion from YCbCr signals to YCbCr signals and image quality improvement processing for display on a monitor.

 It also realizes image processing such as character recognition and facial 'vein' fingerprint recognition.

 [0029] Note that offset correction means 61, gain correction means 62, gradation conversion means 8, and image processing means 9ί, hardware such as ASIC (application-specific integrated circuit), FPGA (field programmable gate array), etc. are used. It can also be configured with software, that is, with a programmed computer.

 [0030] By configuring the image processing means 9 with a personal computer, high-performance embedded microcomputer, DSP (digital 'signal' processor), etc., statistical computation and pattern recognition are performed, thereby eliminating noise components. And complicated signal processing can be easily realized.

 As shown in FIG. 3, the luminance distribution detection means 11 includes a histogram generation unit 111, a maximum level value detection unit 112, and a minimum level value detection unit 113. The maximum level value detection unit 112 and the minimum level value detection unit 113 constitute the maximum / minimum level value detection means 114.

 [0032] The luminance distribution detection means 11 is based on a detection result of a pixel area that is recognized as a subject or a pixel area that is not recognized as a subject, from a subject shape recognition section means 10 (detailed functions will be described later). A luminance distribution of only a pixel region recognized as a subject is detected.

When the signal level of the pixel area that is not the subject is different from the signal level of the subject, for example, when the signal level is lower or higher than the signal level of the subject, the signal level of the subject is detected effectively. And detection accuracy is improved. Note that the luminance distribution detection means 11 can also store image information of one frame or more. By detecting the luminance distribution for each of a plurality of frames, the detection error of the luminance distribution between the frames can be reduced, and the accuracy of the output signal of the luminance distribution detecting means 11 can be improved.

 [0034] The luminance distribution detection means 11 uses the AZD output Sc for one frame from the pixel data of the region recognized as the subject by the subject shape recognition unit 10 and the signal level of the green signal and the luminance signal (Y). The histogram generation unit 111 generates a luminance signal level (gradation) distribution (brightness histogram) according to the signal level, and the maximum level value detection unit 112 and the minimum level value detection unit 113 generate the luminance histogram. Then, obtain the maximum and minimum level values of the effective luminance signal level of the subject.

 Here, “effective luminance signal level” refers to a signal level other than an isolated value (exceptionally large, value, exceptionally small !, value compared to other values) due to pixel defects, noise, and the like.

 The subject shape recognition unit 10 recognizes the shape pattern of the subject and detects an effective subject area. This subject region means a feature region of the subject (a region corresponding to a finger in the case of fingerprint authentication, a palm or finger in the case of vein authentication, and a face in the case of face authentication).

 [0036] The validity of the detection result of the subject area is determined by a geometric characteristic such as a shape unique to the subject, for example, a fingerprint shape, a vein shape, a relative positional relationship between eyes, nose, and mouth. There are a method and a method of judging based on the characteristics of the color and luminance change pattern, but details are omitted.

 [0037] It should be noted that the generation of the luminance histogram and the detection of the maximum and minimum level values of the effective luminance signal level may be performed only within the subject region, or may be performed within the subject region and outside the subject region. However, for the sake of simplicity, the following description will be limited to an example in the subject area only.

[0038] When the subject shape recognition unit 10 recognizes the subject as a subject to the luminance distribution detection unit 11, a signal having a value indicating the recognition, for example, "1"("recognitionsignal" or "in-region signal") If the signal cannot be recognized, a signal indicating that the signal is not recognized, for example, a signal having “0” (“recognition impossible signal” or “out-of-region signal”) is output. The luminance distribution detecting means 11 can determine whether or not the luminance distribution needs to be detected as described above. Note that the output of the subject shape recognition means 10 may control the correction amount determination means 12. In this case, the correction amount can be set in the correction amount determination unit 12 based on the output of the luminance distribution detection unit 11 and the output of the object shape recognition unit 10.

FIG. 4 is a diagram illustrating an example of a histogram generated by the histogram generation unit 111. The horizontal axis shows the gradation, but the gradation from 0 to 1023 of the pixel data Sc is divided into 32 sections, and the average value of the gradation values of each section is represented as a representative value. That is, the histogram generation unit 111 generates a histogram when the frequency is counted by dividing every 32 gradations, that is, 1024 gradations into 32 sections.

[0040] As described above, when 1024 gradations are divided into 32 sections, the number of gradation values per section is 32 as described above. The numbers shown on the horizontal axis in Fig. 4 show the gradation values near the center in the category as representative values. For example, the numerical value “16” on the horizontal axis is a representative value of the division of the gradation value 0 to 3 1 power, and the frequency shown at the position “16” is included in the pixel data Sc of one frame. Represents the number of pixels having gradation values of 0 to 31.

 In addition, when each division consists of one gradation value and the number of gradation values of the pixel data Sc, for example, the pixel data is 10 bits, it is generated for each of 1024 gradation values such as 0 to 1023. You may make it do.

[0041] The histogram generation unit 111 counts the frequency for each gradation division of the pixel data Sc for one frame, and generates a histogram for one frame as shown in FIG.

 The maximum level value detection unit 112 detects a gradation division in which the frequency Has accumulated in the lower (lower) division of the brightest (larger) gradation division power of the generated histogram exceeds a preset threshold Cta, and The representative value of the category is output as the maximum level value Sea. In the same manner, the minimum level value detection unit 113 performs the gradation classification in which the frequency Hbs in which the darkest (smallest) gradation classification power of the generated histogram is sequentially accumulated in the higher classification exceeds the preset threshold value Ctb. Is detected and the representative value of that category is output as the minimum level value Scb.

[0042] By accumulating the frequencies in the descending order of the gradation values in the histogram generated as described above, the cumulative frequency Hbs on the dark side is obtained, and similarly, the classification power having the high gradation values is obtained. By accumulating the frequencies in order, it is possible to obtain the brightness and the accumulated frequency Has on the side. Cumulative frequency Hbs If the threshold value Ctb exceeds the preset threshold value Ctb, the minimum level value Scb (48 in Fig. 4) is set, and if the cumulative frequency Has exceeds the set value Cta, the maximum level value Sea (in case of Fig. 4 84) Output as 8).

 [0043] It should be noted that the configuration of detecting the maximum luminance value and the minimum luminance value from the luminance signal level of the image area whose shape has been recognized as a subject without using a histogram reduces the hardware gate scale. It is possible to shorten the software processing time.

 The maximum level value Sca and the minimum level value Scb are supplied to the correction amount determination means 12.

 [0045] The correction amount determination means 12 obtains the offset correction amount Kb and the gain correction amount Ka based on the maximum level value Sea and the minimum level value Scb supplied from the luminance distribution detection means 11! The correction method by the offset correction means 61 and the gain correction means 62 will be described, and then the method for obtaining Kb and Ka will be described.

 It is assumed that the maximum level value Sea and the minimum level value Scb obtained by the luminance distribution detection means 11 are as shown in FIG. In the circuit of FIG. 1, the AZD output Sc is stored in the imaging signal memory 5 and then read out from the imaging signal memory 5 and supplied to the correcting means 6. The signal supplied to the correction means 6 is the force represented by the symbol Sd, and its value itself is the same as the AZD output Sc. Further, as will be described later, the offset correction amount Kb and the gain correction amount Ka determined based on the AZD output Sc are read after the AZD output Sc is stored in the imaging signal memory 5 and corrected by the correction means 6. Is used for the correction. Therefore, the following description will be made assuming that the AZD output Sc is input to the correction means 6 as it is for a while.

 Further, in FIG. 5, the value of the AZD output Sc and the signal S d read from the imaging signal memory 5 is represented by the symbol X, and the maximum level value Sca and the minimum level value Scb in the one frame are respectively shown. Expressed by Xa and Xb, and for convenience, the output of correction means 6 is expressed by Y, and the maximum level value of the variable range (the range of possible values) is expressed by Ym.

[0048] The function of the correction means 6 is that when the input X is the minimum level value Xb, the value of the corrected imaging signal (output of the correction means 6) Y (= Sf) is α XYm and the input of the correction means 6 When X is the maximum level value Xa, the value of the corrected imaging signal Y is (1−j8) XYm. Here, α and j8 are both set to 0.1, for example.

 In this way, whatever the average value of the input X (i.e., AZD output Sc) of the correction means 6 is, and the difference between the maximum level value Xa and the minimum level value Xb is small, the maximum level is reached. The value Sea and the minimum level value Scb are converted to (1 j8) XYm, a XYm, and the change in luminance is expanded to the range of a XYm force (1—) XYm. It can be performed.

[0049] However, in the present embodiment, correction is performed only when the minimum level value Xb (= Scb) is larger than a predetermined threshold value Xbt (eg, 400), and the minimum level value Xb (= Scb) Is not corrected (set Kb = 0, Ka = l) when is below a predetermined threshold value X bt (eg 400). Even when the minimum level value Xb is equal to or less than a predetermined threshold value Xbt (eg 400), the minimum level value Xb is larger than the threshold value Xb, and correction may be performed in the same manner as that.

[0050] The following relationship exists between Xa, Xb, and Ym.

 (Xa-Kb) XKa = (l- j8) XYm… hi)

 (Xb-Kb) XKa = a XYm

 Solving equations (1) and (2) as simultaneous,

 Kb = {o; XXa— (l—j8) XXb} Z {a— (l—j8)} · '· (3)

 Ka = [a XYmX {a- (l-j8)}] / {«X (Xb—Xa)}

 … (

 If a = β,

 Kb = {a XXa- (l- a) XXb} / {2X α -1)} · '· (5)

 Ka = {α XYmX (2Χ α ~ 1)} / {a X (Xb—Xa)}

 ー (6)

 [0051] In addition,

 If a = j8 = 0.1, then

 Kb = {0. LXXa-O. 9XXb} / {2XO. 1— 1)}

 = {0. lXXa-O. 9XXb} / (-0.8) --- (7)

 Ka = {0. LXYmX (2X0. 1— 1)}

/ {(-O. L) X (Xb-Xa)} = {—0. 08 XYm} / {0. IX (Xb-Xa)}

 ...

 Suppose Xm = Ym,

 Xa = 0.8 XXm, Xb = 0.4 XXm

 Kb = 0.35 XXm "-(9)

 Ka = 2--(10)

 [0052] The correction amount determining means 12 obtains the gain correction amount Ka and the offset correction amount Kb from the equations (3) and (4), and uses the gain correction amount Ka and the offset correction amount Kb used in this way as gains. Supply to correction means 7 and offset correction means 61.

 The offset correction unit 61 subtracts the offset correction amount Kb from the output of the imaging signal memory 5, and the gain correction unit 62 multiplies the output Se of the offset correction unit 61 by the gain correction amount Ka to obtain the gain correction output Y. (= Sf) is output.

 [0053] As described above, subtraction of the offset correction amount Kb in the offset correction unit 61 and correction by multiplication of the gain correction amount Ka in the gain correction unit 62 are the AZD output, that is, the maximum level value Xa of the input of the correction unit 6 Is set to a value (1 β) X Ym that is close to the maximum value Ym of the variable range of the output of the correction means 6, that is, the input of the gradation conversion means 8, and the AZD output, that is, the minimum level value Xb of the input of the correction means 6 is corrected The output of the means 6, ie, the value (α X Ym) close to the minimum value (0) of the variable range of the input of the gradation conversion means 8, and the output of the AZD conversion, ie the maximum level value of the input of the correction means 6 Xa The values other than the minimum level value Xb are values evenly allocated in the range from (1 j8) XYm to (X XYm).

 Note that the gain correction amount Ka may be set to a value smaller than the value obtained by the equation (4) instead of the value of the equation (4).

 Further, the values of the offset correction amount Kb and the gain correction amount Ka are prepared in advance based on the range of the minimum level value Xb and the maximum level value Xa without being obtained by the above equations (3) and (4). You may make it require in the table.

Furthermore, in the above example, the offset correction amount Kb from only the minimum level value Xb is a simple method for obtaining the offset correction amount Kb and the gain correction amount Ka of the minimum level value Xb and the maximum level value Xa. The average value of the maximum level value Xa and the minimum level value Xb (intermediate) Value) and the maximum value Ym of the output variable range of the correction means 21 may be used to determine the gain correction amount Ka.

 Fig. 6 is an example of a table used to determine the gain correction amount Ka from the minimum level value Xb.

 The correction amount determination unit 12 also controls the tone conversion characteristics of the tone conversion unit 8 based on the tone conversion control signal Kc. An example of the operation for controlling the gradation conversion characteristics by the correction amount determining means 12 will be described with reference to FIGS. 7 (A) and (B) and FIGS. 8 (A) and (B).

 FIG. 7 (A) shows a one-frame histogram of the output Sc of the AZD conversion means 3, and a low gradation (in the example shown, the gradation is a gradation value whose representative value is 447 or less). The appearance frequency of 高 shows high ヽ characteristics. For an object in such a state, the gradation characteristic indicated by the broken line TCC1 in FIG. 7B can be used to output with emphasized low gradation contrast.

 FIG. 8A shows a characteristic in which the appearance frequency of the intermediate gradation (in the example shown, the gradation value category representative values are 447 to 831 gradation) is high. For a subject in such a state, the gradation characteristic shown by the broken line TCC2 in FIG.

 As described above, it is possible to determine the gradation conversion control signal Kc based on the histogram of the gradation values of the output Sc of the AZD conversion means 3, and to control the gradation characteristics by the gradation conversion control signal Kc. It will be possible.

 [0060] As a control method, the gradation characteristics realized by the lookup table (shown by broken lines TCC1 and TCC2 in FIGS. 7B and 8B) have several characteristics. In addition, using a logic circuit that can be switched based on Kc, the gradation characteristics composed of broken lines (shown by solid lines TCD1 and TCD2 in Fig. 7 (B) and Fig. 8 (B)) are broken. The coordinate value of the point may be changed based on the gradation conversion control signal Kc.

FIGS. 9A to 9G are timing charts showing the operation of the imaging apparatus of the first embodiment. VD in FIG. 9 (A) indicates a synchronization signal (vertical synchronization signal) in one frame period (Tf). Fig. 9 (B) shows the output timing of the AZD conversion output Sc, Fig. 9 (C) shows the readout timing of the imaging memory output Sd, and Fig. 9 (D) shows the maximum level Xa and the minimum level. The timing for determining Xb and correction amounts Ka and Kb is shown. Fig. 9 (E) shows the output timing of correction amounts Xa and Xb. FIG. 9 (F) shows the output timing of the signal Sf from the correction means 6. Fig. 9 (G) shows the tone conversion characteristics Ft used.

[0062] The frame period can be adjusted according to the imaging conditions such as high-speed reading and long exposure. In the case of high-speed reading, the frame period is set short (high frame rate) and long. In the case of time exposure, a long frame period (low frame rate) is set.

 Note that Ft (a) shown in FIG. 9 (G) is used as the gradation conversion characteristic in the gradation conversion means 8.

 [0064] In the first frame period F1, the AZD output Scl shown in FIG. 9B is output and written to the imaging signal memory means 5, and in the subsequent second frame period F2, the AZD output Scl shown in FIG. As shown in), it is read out as the imaging signal memory output Sdl. The same applies to the following, and the imaging memory output Sdi (i = 1, 2,...) Has the same contents as the AZD output Sci (i = 1, 2,...).

 In the first frame period F1, as shown in FIG. 9B, the AZD output Scl is written to the imaging memory means 5 and supplied to the luminance distribution detecting means 11. The luminance distribution detection means 11 receives the AZD output Scl and generates a histogram. In the last blanking period BL of the first frame period F1, as shown in FIG. 9 (D), as shown in FIG. The detection of the bell value Xbl is performed, and the result is supplied to the correction amount determination means 12. The correction amount determination means 12 determines the first frame period based on the supplied maximum level value Xal and minimum level value Xbl. In F1, the offset correction amount Kbl and gain correction amount Kal for the data Scl output from the AZD conversion means 3 are determined.

 As shown in FIG. 9 (E), the determined offset correction amount Kb 1 and gain correction amount Kal are output from the AZD conversion means 3 in the next second frame period F2 (at that time, the first frame period F1). The same data Sdl as the inputted data Sc 1 is output from the imaging signal memory means 4) and supplied to the offset correction means 61 and the gain correction means 62. As a result of correction using these correction amounts, the signal Sf shown in FIG.

[0066] During the second frame period F2, the memory output Sdl is read from the imaging signal memory means 5 (FIG. 9C), and the imaging memory means 5 of the AZD output Sc2 (the data of the next frame). Writing to (Figure 9 (B)) is performed. [0067] In the same manner, data Sdi (= Xi) having the same contents as the data Sci output from the AZD conversion means 3 in the i-th frame period Fi (i = 1, 2,...) The offset correction amount Kbi and the gain correction amount determined based on the data Sci output from the AZD conversion means 3 during the i-th frame period Fi when read from the means 5 and supplied to the correction means 6 Kai is supplied to correction means 6 for correction.

[0068] The content of this correction is

 Yi = (Xi-Kbi) XKai (i = l, 2,-) ー （11)

 It is represented by

 The corrected imaging signal Sf (= Y) is subjected to the gradation by the gradation conversion characteristic Ft (a) by the gradation conversion means 8, and the gradation conversion output Sg obtained as a result is output to the image processing means 9. .

FIG. 10 is a flowchart showing a processing procedure of the first embodiment.

[0070] First, shooting is performed in step S 1! ヽ, AZ corresponding to the image signal obtained from subject 2

D output Sc is obtained. In step S2, the maximum level value Xa and the minimum level value Xb are detected.

In step S3, it is determined whether or not to perform correction by comparing the minimum level value Xb with a threshold value Xbt (for example, Xbt = 400). If it is greater than the minimum level value Xb threshold value Xb, it is determined that correction is necessary, and the process proceeds to step S4.

In step S4, a gain correction amount Ka and an offset correction amount Kb are calculated. The calculation of the gain correction amount Ka and the offset correction amount Kb is performed by the method described above for the correction amount determination means 12.

 In step S5, correction processing is performed using the gain correction amount Ka and the offset correction amount Kb.

 In step S6, gradation conversion is performed.

 [0074] If it is determined in step S3 that the minimum level value Xb is equal to or smaller than the threshold value Xbt and correction is not necessary, correction is not performed (that is, without being subtracted by the offset correction means 61 (in other words, offset). The gradation conversion (step S6) is performed without increasing the gain by the gain correction means 62 (the gain correction amount Ka is set to “1”) without the correction amount Kb being “0”.

[0075] Among the above, the process of step S1 is performed by the solid-state imaging device 1, the processes of step S2 and step S3 are performed by the luminance distribution detecting means 11, and the process of step S4 is the correction amount determining means. 12 and the processing of step S5 is performed by the correction means 6 and the step The process of S6 is performed by the gradation converting means 8.

 [0076] It should be noted that, for example, if it is detected from the maximum level value Xa, the minimum level value Xb, or other information obtained from the histogram that, for example, the signal amount of the previous frame force has changed extremely (significantly), correction is performed. It is also possible not to perform.

 [0077] The AZD output Sc corresponding to the image signal obtained from the subject is as shown by reference numeral Fa in FIG. 11B, and has a high-frequency component that vibrates with a peak of approximately 1023 and a bottom of approximately 623. Assume that The image signal Fa shown in FIG. 11B is obtained when the subject is generally bright (when the DC component is large) and there are few low-luminance signals.

 In FIG. 11 (A), the gradation conversion means 8 will be described using the case where the γ characteristic in the case of y = 0.7 is used as the conversion characteristic.

 Assuming that the signal Fa is subjected to gradation conversion as it is, the output of the gradation converting means 8 has an amplitude of a higher frequency signal component than the input as indicated by reference numeral Ga in FIG. (Here, the amplitude is compared with a value obtained by normalizing the maximum value in the range of values that the input and output of the gradation conversion means 8 can take as 1).

 [0079] Therefore, offset correction is applied, and the difference between the peak value and bottom value of the high-frequency signal component is maintained, and the signal Fb is formed by shifting to the low luminance side as indicated by the arrow FG. When gradation conversion is performed by inputting to the conversion means 8, the output of the gradation conversion means 8 becomes as indicated by the symbol Gb in FIG. 11C, and the amplitude of the high frequency signal component becomes larger.

 [0080] Further, when the gain conversion is performed only by measuring the offset correction, the signal Fc is formed, and when the gradation conversion is performed by inputting the signal Fc into the gradation conversion unit 8, the gradation conversion unit 8 The output of is as shown by Gc in Fig. 11 (C), and the amplitude of the high-frequency signal component is even greater.

 In this way, after the signal Fa is converted into the signal Fb or Fc by the correcting means 21, and then input to the gradation converting means 8, a high frequency signal component, in other words, is output to the output side of the gradation converting means 8. It is possible to reproduce a fine pattern in detail of an image clearly. Therefore, small brightness differences in the details of the subject image, such as human facial features, blood vessel patterns, and fingerprint patterns, can be reproduced clearly.

[0081] By performing the processing of Embodiment 1, for example, the imaging result of the original image is changed as shown in FIG. Even if it is as shown in (B), the corrected image signal shown in FIG. 12 (C) can be obtained. Here, a description will be added with reference to FIGS. Fig. 12 (A) shows an outline of the original image of the subject, and Fig. 12 (B) shows an image signal along the broken line SL in Fig. 12 (A) (a signal obtained by sequentially reading signals from the pixels on the broken line SL. FIG. 12C shows a signal obtained by applying the correction according to the present embodiment to the signal of FIG. 12B. As shown in Fig. 12 (A), the amount of reflected light differs between the ridge and other parts (the amount of scattered light and the amount of transmitted light may be different depending on the configuration of the optical system). As shown in (B), the signal difference (contrast) with respect to the fingerprint irregularities is output as the signal amount. Also, the center of the finger is lighter than the edge of the finger. When such a subject is imaged, the signal level increases toward the center of the finger force as shown in FIG. 12 (B), and a signal in which the contrast of the fingerprint unevenness signal is superimposed is output.

 By performing the processing of Embodiment 1, it is possible to output a signal in which the contrast of the unevenness of the fingerprint is clear as shown in FIG.

 Figures 12 (A) to 12 (C) are examples, and the same effect can be achieved for signals for which contrast is to be enhanced.

[0082] Embodiment 2.

 The overall configuration of the imaging apparatus of the second embodiment is as shown in FIG. 1, but differs in the following points. That is, in Embodiment 1, the maximum level value and the minimum level value of the pixel signal of one frame are obtained, and based on this, the offset correction amount Kb and the gain correction amount Ka for the pixel data of that frame are obtained, and the same frame is obtained. As shown in Fig. 13, the screen is divided into multiple areas or blocks DV1 to DV12, and the maximum level value and the minimum level value are detected separately for each area, and offset correction is performed. The amount Kb and the gain correction amount Ka may be determined and used to correct the pixel data in each area.

[0083] In order for the offset correction means 61, the gain correction means 62, the luminance distribution detection means 11, and the correction amount determination means 12 to perform processing for each area, the luminance distribution detection means 11 has a histogram for each area. The maximum level value Sea and the minimum level value Scb are obtained, and the correction amount determination means 12 determines the offset correction amount Kb and the gain correction amount Ka for each area. So Then, the correction means 6 corrects the pixel data of the pixels in each area using the offset correction amount Kb and the gain correction amount Ka. The other points are the same as in the first embodiment described with reference to FIG.

FIG. 14 shows a processing procedure of the second embodiment. FIG. 14 is generally the same as FIG. 10, except that processing for each area (step S8, step S9, step S10, step SI 1) is added.

 In step S8, the number of areas is initialized (set to 0). In step S9, the final area force in the screen (the 12th area force in the example shown in FIG. 13) is determined. If NO in step S9, go back to step S2. If YES in step S9, the process proceeds to step S10. In step S10, the number of areas is initialized (set to 0). In step S11, the final area force in the screen (the 12th area force in the example shown in FIG. 13) is determined. If NO in step S11, the process returns to step S3. If YES in step S11, proceed to step S6.

 [0086] The processing of step S8 and step S9 is performed by the luminance distribution detection means 11, and the processing of step S10 and step S11 is performed by the correction amount determination means 12.

 [0087] By dividing the screen into a plurality of areas, it is possible to correct the offset of the brightness for each area, and it is possible to improve the contrast optimum for each area.

 [0088] Embodiment 3.

 The overall configuration of the imaging apparatus of the third embodiment is as shown in FIG. 1, but unlike the first and second embodiments, the imaging signal memory means 5 can be obtained by performing correction for each area. It is possible to use a memory with a memory capacity smaller than one frame, for example, a shift register with several stages (about 2 to 8 stages), a line memory (memory for one line of pixel data), and a FIFO.

 In addition, the luminance distribution detection means 11 uses, for example, line data of pixels of several pixels in the horizontal direction and several pixels in the vertical direction around the pixel at an arbitrary position in the imaging screen of the solid-state imaging device as a line. Read out from memory and create a histogram based on it to detect maximum level value Sca and minimum level value Scb.

In other respects, Embodiment 3 is the same as Embodiment 1 and Embodiment 2. However, the same effect as in the first and second embodiments can be obtained. Further, since there is no need to use frame memory, processing speed can be increased and costs can be reduced.

FIGS. 15A to 15G are timing charts of the third embodiment. In the figure, “PS” shown in FIG. 15 (A) indicates a synchronization signal in the basic unit of time of processing, and the luminance distribution is represented by PI, P2, P3,... Within the processing period divided by this synchronization signal. Detection, gain correction amount Ka, offset correction amount Kb determination, etc. are executed. The processing cycle is a period shorter than one frame period with the horizontal transfer clock of the solid-state imaging device 1 as a minimum unit.

 In addition, as a unit of time for basic processing, a time longer than the time that can realize signal processing of one block is used.

 FIG. 15B shows the output timing of the AZD conversion output Sc, FIG. 15C shows the readout timing of the imaging memory output Sd, and FIG. 15D shows the maximum level Xa Figure 15 (E) shows the timing of output of correction amounts Xa and Xb, and Fig. 15 (F) shows the signal from correction means 6. Indicates the output timing of Sf. Fig. 15 (G) shows the tone conversion characteristics Ft used.

 Note that, in the gradation conversion means 8, the same gradation conversion characteristic (indicated by the symbol Ft (a)) continues to be used.

 In the first processing cycle P1, as shown in FIG. 15B, the AZD output Scl is written into the imaging memory means 5 and supplied to the luminance distribution detection means 11. The luminance distribution detection means 1 1 receives the AZD output Scl and generates a histogram, and in the last blanking period BL of the first processing cycle F1, as shown in FIG. 15 (D), the maximum level value Xal, minimum The level value Xbl is detected, and the result is supplied to the correction amount determination means 12. The correction amount determination means 12 performs the first processing based on the supplied maximum level value Xal and minimum level value Xbl. The offset correction amount Kbl and gain correction amount Kal for the data Scl output from the AZD conversion means 3 are determined in the period P1.

[0094] The determined offset correction amount Kb1 and gain correction amount Kal are obtained by performing AZD conversion on the next second processing cycle P2 (at that time, the first processing cycle P1 as shown in FIG. 15E). The same data Sdl (FIG. 15 (C)) as the data Scl output from the means 3 is supplied to the offset correction means 61 and the gain correction means 62 to the image signal memory means 5). These correction amounts As a result of the correction used, the signal Sf shown in FIG.

[0095] In the next second processing cycle P2, reading of the memory output Sdl from the imaging signal memory means 5 (Fig. 15C) and imaging of the AZD output Sc2 (data of the next processing cycle) are performed. Writing to the memory means 5 (FIG. 15B) is performed.

In the same manner, data Sdi (= Xi) having the same contents as the data Sci output from the AZD conversion means 3 in the i-th processing cycle Pi (i = 1, 2,. The offset correction amount Kbi and the gain correction amount determined based on the data Sci output from the AZD conversion means 3 in the i-th processing cycle Fi when read from 5 and supplied to the correction means 6 Kai is supplied to correction means 6 for correction.

[0097] The content of this correction is

 Yi = (Xi-Kbi) XKai (ί = 1, 2, ...)--(12)

 It is represented by Equation (12) is similar to Equation (11), except that i represents the number of the processing period, not the frame number.

 [0098] Since correction can be realized for each processing cycle Pi, it is possible to realize high-speed detection of small luminance differences in the details of the subject image.

 [0099] Embodiment 4.

 FIG. 16 is a block diagram showing the configuration of the fourth embodiment of the present invention. In FIG. 16, the same reference numerals as those in FIG. 1 denote the same members. The imaging apparatus of the fourth embodiment is generally the same as the imaging apparatus of the first embodiment, but does not use the imaging signal memory means 5 of the first embodiment and does not incorporate the image processing means 9 in the imaging apparatus main body. 2 shows a configuration of an imaging apparatus in which the correction control means 7 and the subject shape recognition means 10 are not built in the imaging apparatus main body. Here, the image processing means 9, the correction control means 7, and the subject shape recognition means 10 are composed of a signal processing device capable of signal processing such as a personal computer and a microcomputer.

 The imaging apparatus shown in FIG. 16 is further different from that of FIG. 16 in that a reverse gradation conversion unit 13 is provided and a luminance distribution detection unit 14 is provided instead of the luminance distribution detection unit 11 of FIG. The correction amount determining means 12, the inverse gradation converting means 13, and the luminance distribution detecting means 14 constitute a correction control means 7.

[0100] Inverse gradation conversion means 13 has gradation conversion characteristics opposite to the gradation conversion characteristics of gradation conversion means 8. It is what you have. That is, the product of the conversion characteristic of the gradation conversion means 8 and the conversion characteristic of the inverse gradation conversion characteristic 13 has linearity. Therefore, the output Sg of the reverse gradation converting means 13 is a signal proportional to the imaging output of the solid-state imaging device 1.

 The luminance distribution detection means 14 receives the output Sh of the reverse gradation conversion means 13 not the output of the AZD conversion means 3, and generates a histogram and detects the maximum level value Sha and the minimum level value Shb. In this case, Sha and Shb are used as the maximum level value Xa and the minimum level value Xb described in the first embodiment.

 The correction amount determination means 12 calculates the gain correction amount Ka and the offset correction amount Kb based on the maximum level value Sha and the minimum level value Shb detected by the luminance distribution detection means 14.

[0101] The signal Sg obtained by gradation conversion by the gradation conversion means 8 is converted into the original gradation characteristic by the inverse gradation conversion means 13.

 Since it is returned to the signal Sh of (characteristic before undergoing gradation conversion), the luminance distribution detecting means 14 and correction amount determining means 12 perform data processing without considering the gradation conversion characteristics of the gradation converting means 8. It can be carried out.

 [0102] In the first embodiment, the luminance distribution detecting means 14 controls correction based on the AZD output (luminance distribution detection, ie, histogram generation, detection of maximum level value, minimum level value, and Ka based on this control). However, the fourth embodiment is different in that the correction is controlled based on the output Sh of the inverse gradation converting means 13 that receives the output Sg of the gradation converting means 8.

 FIG. 17 is a flowchart showing a processing procedure of the fourth embodiment. FIG. 17 is generally the same as FIG. 10 except that step S12 for performing reverse gradation conversion is added. The process of step S12 is performed by the reverse gradation converting means 12.

 FIGS. 18A to 18G are timing charts showing the operation of the imaging apparatus of the fourth embodiment. FIGS. 18A to 18G are the same as FIGS. 9A to 9G. However, FIG. 18 (C) shows the output of the inverse gradation converting means 13. Also, the correction amount Ka (i-1), Kb determined based on the pixel data Sh (i-1) of one frame ((i-1) th frame) F (i-1)! (i 1) is used to correct the pixel data Sci of the next frame (i-th frame) Fi.

 Therefore, the correction contents in the correction means 21 are as follows:

Yi = {Xi-Kb (i- 1)} X Ka (i- 1) (i = l, 2, ...) •••(13)

 It is represented by

 [0105] Generally, when a moving image is captured, the correlation between pixel data of two consecutive frames is high. As shown in FIGS. 18 (C), (D), and (E), moving image correction can be realized by performing correction using correction data determined based on pixel data of the previous frame. . It should be noted that the pixel data Sh (i—1) of one frame ((i−1) th frame) F (i−l) and the pixel data Sh (of the next frame (ith frame) F (i) Based on both i), the correction amounts Ka (i + 1) and Kb (i + 1) of the next frame ((i + 1) th frame) may be determined.

 As described above, by having a signal processing function outside the imaging apparatus main body, it is possible to reduce the cost and mounting area of the imaging apparatus without using a high-performance microcomputer in the imaging apparatus. Further, the correction accuracy by the personal computer can be improved.

 [0107] Note that the fourth embodiment can also be configured to be divided into a plurality of areas and processed as described in the second embodiment.

 [0108] Embodiment 5.

 FIG. 19 is a block diagram showing the configuration of the imaging apparatus according to the fifth embodiment of the present invention. In FIG. 19, the same reference numerals as those in FIG. 16 denote the same members. The image pickup apparatus of the fifth embodiment is generally the same as the image pickup apparatus of the fourth embodiment, but does not use the reverse gradation conversion means 13 of the fourth embodiment, and the gradation conversion means of the fourth embodiment. The gradation conversion means 15 is used instead of 8, the correction amount determination means 16 is used instead of the correction amount determination means 12 of Embodiment 4, and the luminance distribution luminance distribution detection means 14 is replaced with the gradation conversion means 15 The difference is that a histogram is generated in response to the output of, and the maximum level value Sgh and the minimum level value Shb are detected. Furthermore, a conversion characteristic control means 17 for controlling the conversion characteristics of the gradation conversion means 15 is provided. The conversion characteristic control means 17 is also controlled to operate in synchronization with the output from the AZD conversion means 3 (by a control means not shown). In the fifth embodiment, the conversion control means 17, the luminance distribution detection means 14, and the correction amount determination means 16 constitute the correction control means 7.

[0109] The conversion characteristic control means 17 receives the gradation conversion control signal Kc from the correction amount determination means 16 and the frame rate. The tone conversion characteristic of the tone conversion means 15 can be switched between linear conversion and non-linear conversion in units of frames by receiving a signal indicating the timing in units of frames.

[0110] In other respects, the fifth embodiment is the same as the fourth embodiment, and substantially the same effects as the fourth embodiment are obtained.

 FIGS. 20A to 20G are timing charts showing the operation of the imaging apparatus of the fifth embodiment. FIGS. 20A to 20G are the same as FIGS. 18A to 18G. However, it differs in the following points.

 [0112] In the first frame period F1, the correction amount determination unit 16 instructs the conversion characteristic control unit 17 so that the gradation conversion unit 15 performs linear conversion. In Fig. 20 (G), the gradation conversion characteristics when performing linear conversion are expressed as Ft (b). When the gradation conversion means 15 performs linear conversion, the output of the gradation conversion means 15 is a signal proportional to the imaging output of the solid-state imaging device 1. Then, the luminance distribution detecting means 14 generates a histogram based on the output Sgl of the gradation converting means 15, detects the maximum level value Sgal and the minimum level value Sgbl, and the correction amount determining means 16 outputs from the luminance distribution detecting means 14. Based on the maximum level value Sga 1 and the minimum level value Sgb 1, the gain correction amount Kal and the offset correction amount Kbl are determined. In this case, Sgal and Sgbl are used as the maximum level values Xa and Xb in the description of the first embodiment.

 [0113] Only the second frame F2 or several frames after the second frame F2 (for example, the second frame F2 and the third frame F3) are determined based on the data of the first frame. Kal and Kbl are used for correction, and the tone conversion means 15 is represented by a non-linear tone conversion characteristic (reference Ft (a)) selected by the conversion characteristic control means 17 based on the output of the correction amount determination means 16. ) To perform tone conversion.

 As described above, in the fifth embodiment, the gain correction amount Kal and the offset correction amount Kbl determined based on the data of a certain frame (for example, F1) are set only for the next frame (F2) or after the next frame. Used to correct image data of several frames (eg F2, F3).

FIG. 21 is a flowchart showing a processing procedure of the fifth embodiment. FIG. 21 is generally the same as FIG. 10 except that step S14 and step S15 are added. In step S14, a linear gradation conversion characteristic Ft (b) is set as the gradation conversion characteristic. In step S15, a non-linear gradation conversion characteristic Ft (a) is set as the gradation conversion characteristic. The processing of step S14 and step S15 is performed by conversion control means 17. By performing this process, the offset correction amount Kbl and the gain correction amount are obtained based on the data of a certain frame when taking a still image without providing the inverse gradation conversion means 13 of the fourth embodiment. Once Kal has been determined, correction can be performed using the offset correction amount Kbl and gain correction amount Kal over several frames, and gain correction amount Ka and offset correction amount Kb are calculated for each frame. This eliminates the need to perform imaging and shortens the time required for imaging.

 [0117] Embodiment 6.

 In the first to fifth embodiments described above, the luminance distribution detecting means (11, 14) generates a histogram and obtains the minimum level value Xa and the maximum level value Xb. Line, face detection, ear, nose, etc.) is higher than the frequency component of the contrast change of the illumination light, so by detecting the amplitude of the low-frequency component and high-frequency component of the AZD output Sc The local maximum level value and the minimum level value within one frame can be obtained.

 FIG. 22 shows an example of such a luminance distribution detection means 18. FIGS. 23A and 23B are waveform diagrams showing the operation of the luminance distribution means 18 of FIG. Note that the data to be processed is represented by a continuous curve connecting the digital signal values in Fig. 23 (A) and (B), which are digital signals.

 Hereinafter, the case where the luminance distribution detecting means 18 of FIG. 22 is used in place of the luminance distribution detecting means 11 of FIG. 1 will be described. The luminance distribution detecting means 14 of FIG. 22 is the same as the luminance distribution detecting means of FIG. 16 or FIG. It can be used instead of 14.

 The luminance distribution detecting means 18 shown in FIG. 22 includes a low frequency component extracting means 19, a high frequency component extracting means 20, an amplitude detecting means 21, a subtractor 22, and an adder 23.

[0120] The low frequency component extraction means 19 receives the AZD output (indicated by the solid line in Fig. 23 (A)) and extracts the low frequency component (indicated by the dotted line LS in Fig. 23 (A)). This low frequency component can also be viewed as representing the local average of the AZD output Sc. As the low-frequency component extraction means 19, a median filter or epsilon filter can be used in addition to a normal digital LPF. Here, the local average indicates an average of several pixels (for example, horizontal 3 pixels × vertical 3 pixels) and several hundred pixels (for example, horizontal 30 pixels × vertical 30 pixels). [0121] By adjusting the cut-off frequency and frequency response of the low-frequency component extraction means 19, it is possible to detect a luminance change in the local region and a global or overall luminance change of the screen. Changes in luminance due to illumination light can be detected as global or overall luminance changes. By adjusting the cut-off frequency of the low frequency component extraction means 19, the size (area) of the local region can be adjusted, and the luminance change corresponding to the local region size can be detected. In other words, when the illumination light enters the screen as a spot light like a night illumination lamp, or when the whole screen is incident as a uniform illumination like a fluorescent lamp in the room, a change in luminance depending on the situation is detected. can do.

 [0122] The cut-off frequency and frequency response of the low-frequency component extraction means 19 vary from several thousand pixels (for example, horizontal 50 pixels x vertical 50 pixels) to tens of thousands of pixels (for example, horizontal 200 pixels x vertical 200 pixels) Set the frequency so that changes can be detected (pixel pixel x pixel clock).

 The high frequency component extracting means 20 receives the AZD output Sc (or reverse gradation converting means) and extracts the high frequency component (FIG. 23 (B)). As the high-frequency component extraction means 20, a normal digital HPF can be used.

 [0124] The amplitude detection means 21 detects the maximum amplitude (1Z2 of the difference between the peak level value (LP) and the bottom level value (LB)) AM of the local region of the output HS of the high frequency component extraction means 20 The In this case, for example, 1Z2 of the difference between the peak level value of the low frequency component and the bottom level value is set as the maximum amplitude AM.

[0125] Note that the highest and lowest values in the local region may be the peak level value and the bottom level value. Instead, the highest value force is also the nth (n is a natural number) value. The (nth highest, value) may be the peak level value, and the lowest and nth value (nth lowest, value) may be the bottom level value. For this purpose, for example, the amplitude detection means 21 includes a histogram generation unit 211, a peak level value detection unit 212, a bottom level value detection unit 213, and a calculation unit 215 as shown in FIG. A histogram is generated for the data in the above-mentioned local region among the data indicating the high-frequency component HS extracted by the component extraction means 20, and in the same manner as described with reference to FIG. 4 (the maximum level in the description of FIG. 4). Determine the peak level value LP and bottom level value LB in the same way as the value and minimum level value). 1Z2 of the difference between the peak level value LP and the bottom level value LB obtained in this way may be obtained by the computation unit 215, and the result of the computation may be the maximum amplitude AM. Of the above, the peak level value detecting unit 212 and the bottom level value detecting unit 213 constitute a peak / bottom level value detecting unit 214.

 The subtracter 22 subtracts the output AM of the amplitude detector 21 from the output LS of the low frequency component extraction means 19. The subtraction result is used as the minimum level value Xb.

 The adder 23 adds the output LS of the low frequency component extraction means 19 and the output AM of the amplitude detector 21. The addition result is used as the maximum level value Xa.

 The correction amount determination means 12 includes a horizontal counter and a vertical counter, so that the timing at which the imaging output is supplied to the correction means 6 via the imaging signal memory 5 and the offset correction amount Kb in the correction control means 7 The timing of occurrence of the gain correction amount Ka can be correlated.

 [0128] As described above, by using the luminance distribution detecting means 18 of the sixth embodiment, it is possible to correct the local change in average luminance, and the contrast detection result for each area where it is not necessary to specify the area in advance. It is possible to improve the continuity of the image (that is, to prevent a large difference in the contrast detection result between adjacent regions).

For example, as shown in FIG. 25, when there are a bright area A1 and a dark area A2 in the screen, the signal of the bright area A1 is, for example, a force offset as indicated by reference numeral Fa in FIG. After shifting to the lower luminance side by the signal correction amount Kb and converting it to the signal Fb, or adding the gain correction amount Ka to increase the amplitude and converting it to the signal Fc, the gradation conversion is performed. By doing so, high frequency components are reproduced with sufficient contrast. The signal in the dark area A2 is subjected to gradation conversion without offset correction (or only gain correction is performed if necessary!), And is reproduced with sufficient contrast.

[0130] As shown in FIG. 25, it is bright! /, And part of the area A1 has the reason that the area A2 exists. If the illumination is uneven, a part of the subject is not illuminated by the shadow. In such a case, a fine brightness change pattern can be reliably reproduced.

[0131] By performing the processing of the sixth embodiment, for example, the imaging result of the original image is changed as shown in FIG. Even if it is as shown in (B), the corrected image signal shown in FIG. 26 (C) can be obtained. Here, FIGS. 26 (A) and (B) are the same as FIGS. 12 (A) and (B) for the first embodiment. FIG. 26 (C) shows a signal obtained by applying the correction according to the present embodiment to the signal of FIG. 26 (B).

 By performing the processing of Embodiment 6, it is possible to remove the low-frequency luminance component that changes from the dark color that changes from the edge of the finger to the center to the light color, and the contrast of the unevenness of the finger is reduced. A clear signal can be output as shown in 26 (C).

 Figures 26 (A) to 26 (C) are examples, and the same effect can be achieved for signals for which contrast is to be enhanced.

[0132] The correction means, correction control means, gradation conversion means, and image processing means of the above embodiments can be realized at least partially by software, that is, by a programmed computer. Further, the correction means, the correction control means, the gradation conversion means, and the image processing means according to the embodiment of the present invention have been described above, but the correction, correction control, gradation conversion, Image processing methods are also part of this invention.

 Industrial applicability

As an application example of the present invention, gradation correction means suitable for detecting a small change in luminance in the details of the subject, such as characters on a document, human facial features, blood vessel patterns, fingerprint patterns, etc. The present invention can be applied to an imaging apparatus having the above.

Claims

The scope of the claims

 [1] a solid-state image sensor;

 A luminance distribution detecting means for detecting a luminance distribution proportional to an imaging output obtained from the solid-state imaging device and a maximum level and a minimum level thereof;

 A correction unit that performs at least one of offset correction and gain correction on the imaging signal, and a corrected imaging signal output from the correction unit based on the maximum level and the minimum level of the luminance distribution detected by the luminance distribution detection unit An imaging apparatus comprising: a correction amount determining unit that controls a correction amount so as to expand a change range of the image.

 [2] a solid-state image sensor;

 Subject shape recognition means for recognizing the shape of the subject on the screen from the feature amount of the imaging signal proportional to the imaging output obtained from the solid-state imaging device, and outputting a subject region signal indicating the subject region;

 A luminance distribution detecting means for detecting a luminance distribution in the subject area based on the subject area signal output from the subject shape recognizing means;

 Correction means for performing at least one of offset correction and gain correction on the imaging signal; and the correction means based on the maximum level and the minimum level of the luminance distribution in the subject area detected by the luminance distribution detection means! An imaging apparatus comprising: a correction amount determining unit that controls a correction amount so as to expand a change range of the corrected imaging signal in the subject region output from the camera.

[3] The imaging apparatus according to [1], further comprising a gradation conversion unit that performs gradation conversion on an output of the correction unit using a nonlinear conversion characteristic.

[4] The luminance distribution means outputs a signal for operating the correction amount determination means in the case of a pixel region recognized as a subject based on the output of the subject shape recognition means, and is recognized as a subject in advance. 3. The imaging apparatus according to claim 2, wherein, in the case of a pixel region, a signal for inactivating the correction amount determination unit is output.

[5] The nonlinear conversion characteristic of the gradation conversion means is that the gradation conversion means is such that when the input of the gradation conversion means is in the first range, the change in output relative to the change in the input is relatively small. When the input is within the second range, the change in output relative to the change in input is relatively large. The imaging apparatus according to claim 3, wherein the imaging apparatus is a threshold.

 [6] The correction amount determination means determines the maximum level value and the minimum level value within a predetermined range of the signal value proportional to the imaging output of the solid-state imaging device based on the output of the luminance distribution detection means. The input value of the gradation conversion means is converted to a value close to the maximum value and the minimum value of the range of values that can be taken, and values other than the maximum level value and the minimum level value within the predetermined range are converted into values of the gradation conversion means. 4. The imaging apparatus according to claim 3, wherein an offset correction amount and a gain correction amount for determining a value evenly allocated between a maximum value and a minimum value in a range of values that can be input are determined.

 [7] The apparatus further comprises reverse gradation conversion means for converting the output of the gradation conversion means with a conversion characteristic opposite to the conversion characteristic of the gradation conversion means,

 4. The imaging apparatus according to claim 3, wherein the luminance distribution detection unit detects the luminance distribution using the output of the inverse gradation conversion unit as a signal proportional to the imaging output of the solid-state imaging device.

 [8] The luminance distribution means is configured such that the signal level proportional to the output of the solid-state imaging device with respect to the number of pixels of one screen is equal to or greater than a predetermined set value. A signal for operating the correction amount determining means is output, and if the predetermined set value is not satisfied! /, A signal for inactivating the correction amount determining means is output. The imaging apparatus according to 1.

 [9] The luminance distribution detection means includes a histogram generation means for detecting a frequency of occurrence of a signal proportional to the imaging output for each luminance level, and a result detected by the histogram generation means within a predetermined range of the screen. 2. The image pickup apparatus according to claim 1, further comprising maximum / minimum level value detecting means for obtaining a maximum level value and a minimum level value of the luminance level.

 [10] The luminance distribution detecting means includes a signal that is proportional to the imaging output and includes a component representing an image color, and a signal that represents a luminance component of the signal proportional to the imaging output or a green component. 2. The imaging device according to claim 1, wherein the luminance distribution is detected based on the luminance distribution.

[11] In the first state, the gradation conversion means performs gradation conversion with the nonlinear conversion characteristic. In the second state, gradation conversion is performed with linear conversion characteristics.

 The luminance distribution detection unit detects the luminance distribution by using the output of the gradation conversion unit when in the second state as a signal proportional to the imaging output. Imaging device.

 A correction step of receiving a signal proportional to the imaging output of the solid-state imaging device and performing at least one of offset correction and gain correction;

 A correction control step for controlling correction by the correction step;

 A gradation conversion step of gradation-converting a correction output obtained as a result of the correction in the correction step with a nonlinear conversion characteristic;

 The non-linear conversion characteristic of the gradation conversion step is such that when the signal supplied as an input to the gradation conversion step is in the first range, the change in output relative to the change in the input is relatively small. When the signal supplied as input to the conversion step is in the second range, the change in output relative to the change in input is relatively large,

 The correction control step includes

 A luminance distribution detecting step of detecting a luminance distribution from a signal proportional to the imaging output of the solid-state imaging device;

 Based on the detection result in the luminance distribution detection step, if a signal proportional to the imaging output of the solid-state imaging device is supplied to the gradation conversion step as it is, the range located in the first portion When there are a large number of components that fall into the second range, offset correction is performed on the signal proportional to the imaging output of the solid-state imaging element to determine a correction amount for increasing the components that fall within the second range. And a correction amount determination step for

 The correction step corrects a signal proportional to the imaging output of the solid-state imaging device using the correction amount determined in the correction amount determination step.

 An imaging method characterized by the above.