JPH0712192B2 - Image information reader - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to image information by adjusting sensitivity when image information is read in a cycle of photoelectric conversion / accumulation, transfer, holding and reading by an accumulation type photoelectric conversion element. The present invention relates to an image information reading device capable of reading image information having a wide dynamic range and high resolution required for a processing system.

(Technical background of the invention and its problems) Along with the great development of solid-state device technology centered on integrated circuits, image sensors have also made rapid progress, but they are still not satisfactory as such. However, it is necessary to devise a great deal on the side of applying this, and as a result, the image information reading device is considerably complicated and expensive. When reading image information in a wide dynamic range, there are photodiodes, photomultiplier tubes, and the like, but the structure of the scanning mechanism and the like becomes complicated, and not only the device becomes large, but also the total cost of the device increases. There are drawbacks. On the other hand, CCD (Charge Couple
In the case of a reading device using a storage photoelectric conversion element (hereinafter simply referred to as an image sensor) such as a d Device) or a MOS (Metal Oxide Semiconductor), the relative cost with respect to the number of pixels is low, and the scanning is performed electrically. Since it can be carried out automatically, the structure is simple, and the device can be made extremely small, it has been particularly used in recent years.

However, this structure also has the drawback that the dynamic range of the element itself is narrow, and the measures to compensate for this drawback make the reading time extremely long and cause a cost increase on the light source side. There is. For this reason, it has been difficult to use image sensors such as CCDs and MOSs for the purpose of accurately detecting image information such as photographic films having a wide dynamic range from underexposed negatives to overexposed negatives. It was

As a reading method using an image sensor such as a CCD generally used, a method of reading image information in a cycle of photoelectric conversion / accumulation, transfer, holding and reading is used. As a method used for widening the dynamic range on the system by adjusting the sensitivity as described above, the time until the photoelectric conversion / accumulation, transfer, and holding of the above-mentioned read cycle is changed by changing the driving basic clock frequency of the image sensor. By adjusting the sensitivity, the charge storage time during photoelectric conversion is continuously or stepwise changed to adjust the sensitivity to widen the dynamic range in the system.

FIG. 1 shows a general driving / reading system of an image sensor. The image sensor 10 receives a light from an image or the like to perform photoelectric conversion and charge accumulation.
And a holding unit 12 that holds the electric charge accumulated in the photoelectric conversion / accumulation unit 11 after being transferred, and a read register 13 that outputs the electric charge held in the holding unit 12 as an analog image signal PS. ing. Further, the pulse oscillator 1 oscillates a basic clock 4fcp of a predetermined frequency (for example, 6MHz),
A clock signal for driving the image sensor 10 by inputting the basic clock 4fcp to the driving timing unit 2.
Generates CK (φI, φS, φR) and image sensor
A signal indicating the operating state of 10, that is, a pixel signal SP corresponding to one pixel of the image sensor 10 and 1 of the image sensor 10.
Horizontal sync signal Hsync corresponding to line scan and vertical sync signal Vsyn corresponding to one screen scan of image sensor 10
Generate and output c and. The clock signal CK input to the image sensor 10 is, for example, a 4-phase signal φI (φI1 to φI4) that drives the photoelectric conversion / accumulation unit 11, and a holding unit 12.
For driving four-phase signals φS (φS1 to φS4)
And a phase signal φR (φR1 to φR4) of, for example, four phases for driving the read register 13, which have the same frequency (for example, 1.5 MHz) obtained by dividing the basic clock 4fcp. Phase signal (φI1 to φI4, φS1 to φS4, φR1
~ ΦR4) are all out of phase due to a predetermined relationship. Image signal PS read from the image sensor 10
Is converted into a digital true value PSD by the A / D converter 21 in the arithmetic processing unit 20, and the true value PSD is logarithmically converted by the logarithmic converter 22 and stored in the memory 23. In addition, the arithmetic processing unit 20 includes a pixel signal S from the driving timing unit 2.
P, the horizontal synchronizing signal Hsync, and the vertical synchronizing signal Vsync are input, and arithmetic processing according to the operating state of the image sensor 10 is performed.

In such a structure, normally, as shown in FIG.
Photoelectric conversion / accumulation of input light in synchronization with clock signal CK (I
A) and the stored charge is transferred, held and read (RA). Then, when the frequency of the clock signal CK is reduced to, for example, 1/2 in order to widen the dynamic range of the system, the photoelectric conversion / accumulation (IA) becomes correspondingly long as shown in FIG. Sensitivity 1 in Figure 3
You can increase the sensitivity by adjusting the sensitivity from the range R1 to R2. However, correspondingly, the time of transfer, holding and reading (RA) becomes long, and there is a problem that dark current increases and smear (charge mixture during transfer) occurs during transfer and holding. Note that FIG. 3 shows the dynamic range FDR required for a high-density image and the actual dynamic range corresponding to the low-density image of the image sensor 10 itself.
It shows the relationship of CDR, and shows how to move from range R1 to range R2 by changing the accumulation time and obtain a wider overall dynamic range SDR (R1 + R2) as a system. Further, if the frequency of the clock signal CK is further lowered in order to widen the dynamic range, the sensitivity of the image sensor 10 is further increased, but there is also a problem that it takes a lot of time to read the entire image information. is there.

(Object of the Invention) The present invention has been made under the circumstances described above.
An object of the present invention is to provide a high-resolution image reading apparatus which has a relatively high processing speed, can accurately read an image, and has a dynamic range necessary for image information processing.

(Summary of the Invention) The present invention relates to an image information reading device for reading image information by using a storage type photoelectric conversion element that operates in a cycle of photoelectric conversion / storage, transfer, holding and reading. And a storage time changing means for continuously and repeatedly changing the storage time by a set number of clocks; a step for increasing the storage time is determined in advance, and each table is configured in a one-to-one correspondence with the storage time for each step. The storage time Tn at the step n is given by the following equation Tn = TB × a ^n, where TB is the basic storage time, a is set by a preset storage time coefficient, and the storage type photoelectric conversion at the storage time Tn is performed. A logarithmic conversion table in which density values corresponding to A / D converted values of the output signals from the elements are preset; and a table corresponding to the step n is switched, and the A / D converted table is switched from the switched table. Density information reading means for reading a density value with respect to a value; first control for controlling the accumulation time by repeating the accumulation time changing means in accordance with the set number of clocks when performing the photoelectric conversion / accumulation Means and; second control means for controlling the table switching / reading operation of the density information reading means.

Another aspect of the present invention is a storage time changing means that continuously and repeatedly changes only the photoelectric conversion / storage mode for a set number of clocks to change the storage time; a step for increasing the storage time is determined in advance, and storage for each step is performed. Each table is configured in a one-to-one correspondence with time, and the accumulation time Tn in the step n is
Where Tn = TB × a ⁿ where TB is the basic storage time and a is the preset storage time coefficient, and A / D conversion of the output signal from the storage photoelectric conversion element at the storage time Tn is performed. A logarithmic conversion table in which density values corresponding to the values are preset; density information reading means for switching to a table corresponding to the step n and for reading density values corresponding to the A / D converted values from the switched table; When photoelectric conversion / accumulation is performed, first control means for controlling the accumulation time by sequentially changing the set number of clocks and sequentially repeating the accumulation time changing means; and a table of the concentration information reading means. Second control means for controlling the switching / reading operation; and image information synthesizing means for synthesizing the density values sequentially read by the density information reading means by combining them in pixel units Those were.

(Embodiment of the Invention) FIG. 4 is a block diagram corresponding to FIG. 1 showing an embodiment of the image information reading apparatus of the present invention. The phase signal φI output from the driving timing section 2 is shown in FIG. (ΦI1 to φ
I4) is applied to the photoelectric conversion / accumulation unit 11 of the image sensor 10 through the gate circuit 4, and the gate circuit 4 is connected to CP.
The control signal CS from the control means 3 such as U is used for control. Further, the control means 3 is connected to the arithmetic processing section 20, and the pixel signal SP, the horizontal synchronizing signal Hsync, and the vertical synchronizing signal.
The operating state of the image sensor 10 is grasped based on Vsync, and image information can be processed. Therefore, the control means 3 can output the vertical synchronizing signal Vsync from the driving timing section 2, that is, the control signal CS in synchronization with the scanning of one screen. Further, the control signal is sent from the control means 3 to the logarithmic table circuit 24 in the arithmetic processing section 20.
A selection signal SL corresponding to CS is output.

The contents of the logarithmic table in the logarithmic table circuit 24 composed of a ROM (Reed Only Memory) or the like will now be described. The relationship between the true value Y and the density value X is as shown in FIG. 8-bit output from A / D converter 21 (0 to 255)
When the density resolution is 0.01, the range of density 0.00 to 0.77 is the effective area of density resolution 0.01 in Table # 0, and the range of density 0.51 to 1.32 is the effective area of density resolution 0.01 in Table # 5. Furthermore, in Table # 10, the density is 1.
The range of 03 to 1.92 is the range of density resolution 0.01. Therefore, if a predetermined number of such effective area tables are prepared over the required density range, the 8-bit A / D converter 21
Is saturated (225). All the range up to the density value 2.55 is 0.0
With a high resolution of 1, the density value X can be accurately converted with almost no influence of noise components such as dark current and offset. That is, the dotted line portion in FIG. 5 has extremely poor resolution as compared with the solid line portion, and accuracy cannot be guaranteed when performing digital arithmetic processing. However, for example, the solid line portions of Tables # 0 and # 5 are appropriately combined. In this case, even if the dynamic range of the image sensor is D = 1.0 (10: 1) or less, the range of density values D = 0.00 to 1.32 can be read with a resolution of 0.01. In this case, in FIG. 3, R1 of the dynamic range SDR (= R1 + R2) corresponds to table # 0 and R2 corresponds to table # 5, and the range R1 and range R2 are combined to have a wide dynamic range and high resolution. Can be obtained.

In such a configuration, the basic clock 4fcp from the pulse oscillator 1 is input to the driving timing unit 2, and the clock signal CK, the pixel signal SP, and the horizontal synchronization signal Hsy are input as described above.
nc and the state signal of the vertical synchronizing signal Vsync, the phase signals φS and φR of the clock signal CK are directly applied to the holding unit 12 and the read register 13 of the image sensor 10, respectively, and the phase signal φI is applied to the gate circuit 4. Via photoelectric conversion
It is applied to the storage unit 11. The image signal PS from the image sensor 10 is input to the arithmetic processing unit 20 and processed, which is exactly the same as described above. Here, the control means 3 is an arithmetic processing unit.
The operating state of the image sensor 10, that is, the cycle mode of photoelectric conversion / accumulation, transfer, holding and reading is determined via 20 and the control circuit CS is switched to control the gate circuit 4. This can also be performed by directly inputting the status signals (SP, Hsync, Vsync) from the drive timing section 2 to the control means 3. In this way, the control means 3 detects the photoelectric conversion / accumulation mode of the image sensor 10, and when the gate circuit 4 is switched by the control signal CS, the phase signals φI1 to φI4 from the gate circuit 4 are logic "L" or "H". For a given combination,
For example, φI1 = "L", φI2 = "L", φI3 = "H", φI4 = "H"
And is provided to the photoelectric conversion / accumulation unit 11. In this case, the phase signals φS and φR are input to the holding unit 12 and the read register 13, respectively. A vertical synchronizing signal output corresponding to one screen scan by the fixed operation time of the phase signals φI1 to φI4 from the gate circuit 4 by the control signal CS.
If it is performed in synchronization with Vsync, only the photoelectric conversion / storage mode is performed multiple times (two times in this example) as shown in FIG. 2 (C).
Can be repeated only. That is, when the control means 3 determines that the image sensor 10 is in the photoelectric conversion / accumulation mode (time point t1), the control means 3 gives the control signal CS to the gate circuit 4 to apply the phase signals φI1 to φI4 to predetermined logic levels. Fixed to the combination of, and photoelectric conversion / accumulation is performed. Then, when photoelectric conversion / accumulation is performed a plurality of times in synchronization with the vertical synchronization signal Vsync, the control means 3 erases the control signal CS to restore the gate circuit 4 (time point t3), and the driving timing unit 2 outputs the signal. The phase signal φI is applied to the photoelectric conversion / accumulation unit 11 as it is. As a result, the image sensor 10
The accumulated charge is transferred, held, and read from t3, and the next operation starts from time t4 when the next vertical synchronization signal Vsync is input.

In the present invention, the control means 3 selects and uses the logarithmic table in the logarithmic table circuit 24 by the selection signal SL according to the control of the gate circuit 4 by the control signal CS.

First, a method of setting the above logarithmic table will be described.
Here, the logarithm is a common logarithm whose base is "10", the basic storage time of the image sensor 10 is TB, the photometric storage time is TX, and the scan (hereinafter referred to as the main scan) for photometry using the logarithmic table is performed. A / D conversion value (exact value) is Y, photometric density value from logarithmic table circuit 24 is X, photometric brightness value is P, accumulation time coefficient is a, logarithmic table number is Tn, density coefficient is K, logarithmic table The number (page) is n, the logarithm conversion table using the antilogarithm table is the maximum antilogarithmic A / D conversion value during photometry (hereinafter referred to as pre-scan) for selecting the number, YP, and base brightness A / D. Let PB be the reference value and D be the required dynamic range. Since the dynamic range is D and the number of logarithmic tables is Tn, the accumulation time coefficient a is Is defined as

Then, the basic accumulation time TB is set before the photometry of the original film and at the time of the calibration operation for detecting the calibration data with the reference film. In order to set the normal film base to zero reference density and increase the resolution of image information, first, the base brightness PB is measured. At this time, the A / D conversion value with a slight margin with respect to the true number saturation output M of the A / D converter 21 is (M
-Α), the basic storage time TB corresponding to the base brightness PB is selected by sequentially extending the storage time from the minimum storage time at which the image sensor can compose image information.

Next, if necessary, set the photometric accumulation time TX by prescan, but use the exact number table for the original image film you want to measure and use the basic accumulation time TB to measure the A / D.
The photometric accumulation time TX is determined by n determined in the prescan table using the maximum luminance value YP of the true output Y of the converter 21 as address information. That is, TX = TB · a ⁿ (2). The A / D conversion value YP measured and determined by the basic accumulation time TB in (2) above is expressed as YP = PB / a ^n. By converting this equation, logYP = log (PB / a ⁿ ) logYP = Since logPB−n · loga (3), n · loga = logPB−logYP (4), and n = (logPB−logYP) / loga (5). It should be noted that n is calculated by rounding down the decimal places. Therefore, the logarithmic table number n selected by the prescan table memory obtained by the above equation (5) is determined using the A / D conversion maximum luminance value YP at the prescan as the address information.

On the other hand, assuming that the photometric brightness value is P, the A / D conversion value Y at the time of the main scan is Y = P × a ⁿ (6), and the photometric density value X is the common logarithm of the reciprocal of the photoluminance ratio. Therefore, the relationship between the AD reference value PB of the base luminance and the photometric density value X with respect to the photometric luminance value P is defined as X = K · logPB / P (7). Since the equation (6) is converted into P = Y / a ⁿ , substituting this yields the above equation (7) as X = K · log (PB / Y / a ⁿ ) = K · log (PB · a ⁿ / Y) = K · log ( Y / PB · a n) -1 = -K [logY-logPB-n · loga] = K [logPB-logY + n · loga] ...... (8) are obtained, X = K・ LogPB ＋ n ・ K ・ loga-K ・ logY …… (9)
Becomes Therefore, the density photometric value X selected by the logarithmic table memory obtained from the above equation (9) is determined using the logarithmic table number n determined during the prescan and the A / D conversion value Y during the main scan as address information. .

From the above, the configuration of the logarithmic table 24 is as shown in FIG. 6, 29 logarithmic tables # 0 to # 28 are prepared, and the prescan table 241 and the input and the output are output one to one. An exact number table 242 is prepared. In the case of 8-bit processing, the address is 0 to 255, the photometric data is in the range of 0 to 255, and the prescan table 221 selects the table number n by n = (log250-logYP) /1.269 (10). To do. In this case, if the required dynamic range D is set to 1: 1000, the accumulation time coefficient a is calculated from the equation (1). And the above equation (10) is obtained. Also, the logarithmic tables # 0 to # 28 are each composed of 256 bytes, and the density value X is X = 100 · log250 + n · 100 · log (1.269) −100 · logY.
.. (11), the density value X is read at the corresponding addresses 0 to 255 of the logarithmic tables # 0 to # 28. In this case, if the density value D = 0.01 corresponds to “1” of the A / D conversion output value, K = 1 / 0.01 = 100, and the density coefficient K is determined by the balance between the required resolution and the dynamic range. . Further, in the case of 8-bit processing, the density value X is clipped by "255", the fractional part is cut off, and when Y = 0, X = 255.

The logarithmic table set in the logarithmic conversion table circuit 24 as described above is selected by the selection signal SL from the control means 3. Therefore, the image signal PS from the image sensor 10 is converted into the logarithmic table density value X corresponding to the accumulation time.

In this way, the image information PS from the image sensor 10 is subjected to sensitivity adjustment by repeating only charge accumulation several times during photoelectric conversion, and by switching the conversion table corresponding to the output signal in correspondence with the set number of repetitions, for example, A dynamic range FDR range required for a high-density image as shown in FIG. 3 is selected and made compatible, and as a result, the system has a wide dynamic range of SDR. Also, the sensitivity is adjusted by sequentially updating the set number of times
By sequentially updating the conversion table corresponding to the output signal and combining and combining the image information on a pixel-by-pixel basis, for example, the total dynamic range SDR shown in FIG. You will be able to read. In the above, the vertical sync signal Vsync is 2
The photoelectric conversion / accumulation mode is repeated for two scans of a screen, but the accumulation time can be adjusted by setting it to any number of times, which allows the dynamic range on the system to be changed freely. It is possible. That is, the sensitivity of the image sensor 10 is increased in proportion to an integral multiple of the basic time required to scan one screen by controlling the storage time of the image sensor 10 in synchronization with the vertical sync signal Vsync. It is possible to improve sensitivity and easily achieve a wide dynamic range.

The accumulation time of the image sensor 10 may be set independently regardless of the basic time, or the photoelectric conversion / accumulation mode may be continuously performed for the set number of clocks.

Further, in the above description, the drive signal (φI to φ) of the image sensor
Although R) has four phases, it can be applied to an image sensor having an arbitrary number of phases. Further, although the phase signals φS and φR are input to the holding unit and the reading register respectively during the photoelectric conversion / accumulation as described above, a similar gate circuit is provided to fix each phase signal to a predetermined logic level. The operation may be stopped completely.

(Effect of the Invention) As described above, according to the present invention, when image information is read using the storage photoelectric conversion element, only the photoelectric conversion / storage mode can be continuously repeated for a set number of clocks, so that the transfer is performed. The storage time can be extended without changing the holding and reading times. Therefore, it is possible to read highly sensitive image information at a relatively high processing speed. Further, since the storage time can be adjusted according to the set number of clocks, the dynamic range on the system can be freely changed.

In addition, logarithmic conversion from the A / D converted value to the density value of the output signal from the storage photoelectric conversion element is performed by selecting a logarithmic conversion table corresponding to the storage time at the set number of clocks, and previously selecting the selected table. Since it can be performed only by reading the set density information after the logarithmic conversion, the logarithmic converter is not necessary and the arithmetic processing of the logarithmic conversion can be eliminated.
Furthermore, sensitivity is adjusted by sequentially changing the number of set clocks,
Since the image information obtained from the logarithmic conversion table corresponding to the storage time at the set number of clocks can be combined and combined on a pixel-by-pixel basis, it is possible to read image information with a wide dynamic range and high resolution. .

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a drive system of a conventional image sensor, FIGS. 2A to 2C are diagrams for explaining the operation of the image sensor, and FIG. 3 is a diagram for explaining a dynamic range. FIG. 4 is a block diagram showing an example of the drive system of the image sensor according to the present invention, and FIGS. 5 and 6 are views for explaining the logarithmic table. 1 ... Pulse oscillator, 2 ... Drive timing section, 3 ...
... Control means, 4 ... Gate circuit, 10 ... Image sensor, 20 ... Arithmetic processing section.

Claims (2)

[Claims] 1. An image information reading apparatus for reading image information using a storage type photoelectric conversion element which operates in a cycle of photoelectric conversion / storage, transfer, holding and reading, wherein only the photoelectric conversion / storage mode is set a predetermined number of times. Storage time changing means for continuously and repeatedly changing the storage time; a step for increasing the storage time is determined in advance, and each table is configured in a one-to-one correspondence with the storage time for each step. Storage time Tn of the following equation Tn = TB × a ^n, where TB is the basic storage time, a is set by a preset storage time coefficient, and the output signal from the storage type photoelectric conversion element at the storage time Tn is A logarithmic conversion table in which density values corresponding to A / D converted values are preset; and a table corresponding to the step n is switched, and the density values corresponding to the A / D converted values are read from the switched table. Concentration information read-out means for outputting; and first control means for controlling the accumulation time by repeating the accumulation time changing means in accordance with the set number of clocks when performing the photoelectric conversion / accumulation; An image information reading apparatus comprising: a second control means for controlling a table switching / reading operation of the reading means. 2. An image information reading device for reading image information by using a storage type photoelectric conversion element which operates in a cycle of photoelectric conversion / storage, transfer, holding and reading, wherein only the photoelectric conversion / storage mode is set a predetermined number of times. Storage time changing means for continuously and repeatedly changing the storage time; a step for increasing the storage time is determined in advance, and each table is configured in a one-to-one correspondence with the storage time for each step. Storage time Tn of the following equation Tn = TB × a ^n, where TB is the basic storage time, a is set by a preset storage time coefficient, and the output signal from the storage type photoelectric conversion element at the storage time Tn is A logarithmic conversion table in which density values corresponding to A / D converted values are preset; and a table corresponding to the step n is switched, and the density values corresponding to the A / D converted values are read from the switched table. And a first control means for controlling the accumulation time by sequentially changing the set number of clocks and sequentially repeating the accumulation time changing means when performing the photoelectric conversion / accumulation. Second control means for controlling the table switching / reading operation of the density information reading means; and image information synthesizing means for synthesizing by combining pixel values of the density values sequentially read by the density information reading means. An image information reading device characterized by being provided.