GB2155730A

GB2155730A - Picture input equipment

Info

Publication number: GB2155730A
Application number: GB08504801A
Authority: GB
Inventors: Takashi Sakamoto
Original assignee: Dainippon Screen Manufacturing Co Ltd
Current assignee: Dainippon Screen Manufacturing Co Ltd
Priority date: 1984-03-08
Filing date: 1985-02-25
Publication date: 1985-09-25
Also published as: GB2155730B; DE3507933A1; JPS60189372A; GB8504801D0

Abstract

The signals from the cells of a CCD sensor array, in e.g. a camera (1), are individually corrected by dark and full white reference signals. The dark signals for each cell is read and stored in memory (5). Later, when white reference signals or picture signals are output by camera (1), the stored dark signals are subtracted by circuit (4) and the resultant signal is converted to its logarithm by circuit (7). The white reference signals, after logarithmic transformation, are stored in memory (12), ready for subtraction from the logarithmic picture signals in circuit (10). If the dark signals are small, their average may be subtracted from the signals from all cells, making memory (5) redundant. <IMAGE>

Description

SPECIFICATION Picture Input Equipment This invention relates to a picture input equipmentfor reading in with abundant graduation and good fidelity a varied-density picture having a wide variety of densities such as color film or the like by means of a CCD (charge coupled device) array sensor in which a plurality of photoelectric conversion cells of the COD type are arranged either iinearly ortwo-dimensionally.

When reading in an original picture by means of a COD array sensor, it is necessary to correct output signals because the photoelectric conversion characteristics of one cell are different from another cell.

In a conventional picture input equipment, each input image is corrected as disclosed for example in Japanese Utility Model Laid-open No. 19566/1983, namely, by storing beforehand data pertaining to the light sensitivity of respective cells and when reading in an actual picture, multiplying picture signals from the respective cells with the signals which pertain to the light sensitivity of their corresponding cells.

In the above correction method, correction coefficients are determined by taking both lighting conditions and conversion characteristics of the cells into parallel consideration. Accordingly, the input of each picture is not affected by variations in light sensitivity of the individual cells under lighting conditions of high iiluminance equal to that employed when the correction data were stored.

When reading in a varied-density picture having a wide range of densities such as color film, the above correction method develops a shortcoming that because there are differences in dark current among the individual cells of the sensor, no accurate picture read-in can be performed, especially, at a dark part of the original if the differences in sensitivity among the individual cells under high illuminance are solely corrected.

In orderto perform an integrating operation for correction, D/A converter and an analog multiplier or high-speed digital multiplier are required. Even when the accuracy of correction is set for example at 1% or so, the memory capacity required for the multiplication with a 3-digit correction coefficient such as 1.10 or 0.95 is 7 bits (0 to 27-1).

Correspondingly, the circuit structure becomes complex and hence costly, thereby imposing a limitation on the accuracy of correction.

An object of this invention is to provide an optimum picture input equipment which can correct the sensitivity characteristics of the individual cells of a photosensor of the COD type.

In one aspect of this invention, there is thus provided a picture input equipment which comprises: a COD array sensor adapted to perform photoelectric conversion based on the quantities of incident light to respective cell; a first arithmetic circuit for subtracting dark output fractions from corresponding signals obtained from the array sensor; a logarithmic transformation circuit for transforming signals output from the first arithmetic circuit into logarithmic values; memory means for storing logarithmicallytransformed reference signals on the basis of signals output respectively from the cells when incident standard light of white level is input to the array sensor; and a second arithmetic circuit for correcting picture signals input from the cells of the array sensor and subjected to logarithmic transformation by either adding or subtracting white level reference signals, which are read out respectively for the corresponding cells from the memory means, to or from the picture signals.

Owing to the above construction, the present invention permits correction of the conversion characteristics of the individual cells of the COD array sensor respectively in the dark and upon input of white level standard incident light. Accordingly, the accuracy of correction of input picture signals is enhanced and at the same time, the white level reference signals are subjected to logarithmic transformation for their correction through addition or subtraction. Hence, the present invention can bring about numerous effects such that the operation speed and reliability of correction are both improved and the structure of the picture input equipment is relatively simple.

In addition, conditions for such correction can be preset prior to scanning an original. This feature can exhibit significant effects especialiy against changes of various parameters, on which the conversion characteristics depend, along the passage of time.

The above and other objects, features and advantages of the present invention will become apparent from the following description and the appended claims, taken in conjunction with the accompanying drawings, in which: Figure 1 is a circuit diagram showing a picture input equipment according to the first embodiment ofthis invention; Figure 2 is a circuit diagram showing a picture input equipment according to the second embodiment of this invention; Figure 3 is a circuit diagram showing a picture input equipment according to the third embodiment of this invention; and Figure 4 is a characteristic diagram showing the relationship between the incident light quantity e to a COD sensor and the output signals vfrom the COD sensor.

The present invention can be applied to a picture input equipment in which a solid-state image sensor composed of a plurality of photoelectric transducers, namely, cells arranged in an array is formed into a COD (charge coupled device).

A COD sensor is considered in principle to output electrical signals in proportion to an integrated light quantity during a single scanning period. As illustrated in the characteristic diagram of the incident light quantity e vs. the output signals v of Figure 4, the output signals v of the respective cells however differ slightly from one another because the sensitivity a1 and offset value b, of one cell are different a little from those of another cell.

However, the linearity of conversion characteristics is generally maintained. Hence, the relationship between e and V can be expressed by the following equation: v1=a1 - e+bl (1) When actually inputting light, which corresponding to respective picture elements, to the individual cells to obtain output signals, lighting conditions such as illuminance of the original picture perse and illuminanoe distribution, the cos48 law and eciipse phenomenon inherent to the optical image formation system, etc. are added to parameters pertaining to the incident light quantity e, whereby making the matter very complex.

Therefore, taking parameters inherent to each cell into consideration, the incident light quantity e1 can be expressed by the following equation: ei=li * cos40jv T (2) where the the illuminance of original-illuminating light at a position corresponding to the Xth cell Sj; Ri: the reflectivity or transmittance of the original at the same position; 0,: the angle between a line extending from the position corresponding to the cell S1 to the lens center af the optical axis; T: the integrating time of light quantity of the cell Si.

Solving the above equations (1) and (2) with respect to the reflectivity or transmittance R1 the determination of which is now intended, the following equations are derived.

v1=a1 Ii Rj Cos48, T+b, C1 ' R1+b1 (3) where can - cos461 T.

Namely, ci is a value which may be considered as an overall gamma characteristic value reflecting both lighting conditions and integrating time. By this value C and dark output bj, the output characteristics of each cell are determined.

Especially, the integrating time T is the same parameter as the scanning time (charge transferring time) of the CCD sensor. Therefore, even in the case of a picture input method making use of varied scanning time, the picture input equipment according to this invention can cope with each scanning speed, in other words, each of various charge transferring time which the CCD sensor may take.

Although the above equations (1), (2) and (3) represent one-dimensional scanning for the sake of simplification of description, the same equations can be applied to two-dimensional scanning provided that each subscript "i" is replaced by a subscript "jj".

In order to achieve accurate input of light corresponding to the picture elements of each original varied-density picture without altering the various densities of the picture but while suppressing occurrence of variations in light sensitivity, the present invention allows, as mentioned above, the output signal vi to depend only on the reflectivity or transmittance R, of the original even if q and b1 vary in the equation (3).

Referring now to Figure 1 of the accompanying drawings, a camera 1 equipped with a CCD sensor is connected to a change-over switch 2. Each output signal vi of the camera 1 is output via the change over switch 2 eitherto an A/D (analog-to-digital) converter 3 or to the (+) input terminal of an analog addition/subtraction circuit 4.

When storing a dark output fraction, the changeover switch 2 is turned to the A/D converter 3 and the resulting output of the A/D converter 3 is fed to a memory 5, where the dark output fraction b1 is stored using the position (cell number)iof its corresponding cell Si, which will be described herein, as an address.

After the dark output fraction has been stored, the change-over switch 2 is turned to the analog addition/subtraction circuit 4.

Each output signal from the memory 5 is input via the D/A (digital-to-analog) converter 6 to the (-) input terminal of the addition/subtraction circuit 4.

On the other hand, the picture signal vi output from the corresponding cell S, as a result of read-in of the original picture by the camera 1 in synchronization with the input of the output signal is fed via the change-over switch 2 to the (+) input terminal of the addition/subtraction circuit 4.

As a result, a signal v1-b1 which has been obtained by subtracting the dark output fraction bi from the output picture signal vi of the corresponding cell Si is output from the addition/ subtraction circuit 4.

Each output signal of the addition/subtraction circuit 4 is fed to a change-over switch 8 by way of a logarithmic transformation circuit 7. The corresponding logarithmically-transformed output signal log (v,-b,) is output either to the (+) input terminal of an addition/subtraction circuit 9 or to the (+) input terminal of an addition/subtraction circuit 10.

When inputting a signal corresponding to the white level standard incident light through the camera 1, the switch 8 is turned to the addition/ subtraction circuit 9.

In order to take out only the varied fraction of the input signal, the addition/subtraction circuit 9 is fed through the (-) input terminal with an input signal E of a constant value which corresponds to anon- varied value. Then, a signal log (v1-b1)-E is output to the output terminal of the addition/subtraction circuit 9. The signal log (v-b)-E is then delivered via an ND converting circuit 11 to a memory 12.

The reference signal is subjected to logarithmic transformation at the logarithmic transformation circuit 7 on the basis of a signal output from the corresponding cell Si when the white level standard incident light has entered the camera 1. Supposing now the entire surface of the original picture has a reflectivity R,=1 and introducing the equation (3), the following equation can be derived.

log (V,-b1)= log c,R, = log q+ log R1 = log ci (cf. log 1=0) The memory 12 is adapted to store a value log c-E which has been obtained by subtracting the constant value E from log q at the addition/ subtraction circuit 9. The output terminal of the memory 12 is connected via a D/A converter 13 to the (-) input terminal of the addition/subtraction circuit 10.

The above-mentioned operation is preparatory work which is required for the effective operation of the equipment of this invention. When inputting picture information R1 on each picture element of the original through the camera 1 to perform the intended objective, the change-over switch 2 is kept connected to the addition/subtraction circuit 4 whereas the change-over switch 8 is turned to the addition/subtraction circuit 10.

At the addition/subtraction circuit 10, an operation is performed between the signal log (v-b) fed to the(+) inputterminal and the constantvalueEfed to the (-) input terminal and the output signal log c-E read out from the memory 12 via the D/A converter 13. As apparent from the equation (3), the signal fed to the (+) input terminal is expressed by the following equation: log (v,b)=log C1 R1 =log clog R, (4) On the other hand, the signal input to the (-) input terminal is log c-E+E=log c. Therefore, the fraction ci is subtracted from a signal to be output from the output terminal of the addition/subtraction circuit 10, thereby outputting a signal log R.

The overall operation of the equipment of the above embodiment and the correcting operation of input picture signals will next be described briefly.

The correcting operation of input picture signals is effected in the following three steps.

First step: Each dark output fraction bi is stored in this step.

The camera 1 is provided for example with a shutter. Signals output under total light-shield conditions are fed via the switch 2 to the A/D converter 3. Each dark output fraction b, of the corresponding cell Si of the camera 1 is stored as a digital value in the memory 5.

Since the quantity of incident light to the camera 1 is O (zero), the dark output fraction is v=bi as apparent from the equation (1). As shown in Figure 4, variations of this value bare small. Hence, the accuracy of the A/D converter 3 is determined depending how many times of variation would occur relative to a desired dynamic range.

If the dynamic range is 1000:1 and b1 is always below 15 for example, it is sufficient for the ND converter 3 to have an accuracy level of 4 bits or so.

It is also understood to be sufficient if the memory 5 contains 4 bits for each sensor.

The dark output fraction b1 is inherent to each cell.

Where a CCD sensor is a line sensor, the dark output fraction should be obtained by scanning cells arranged on a single line. Where a CCD sensor is an area sensor, it should be obtained by scanning all cells.

Second step: This is a step for setting up a white level standard (i.e., a step for setting up a correction value for the white level).

The while level standard incident light is input to the camera 1 under white paper conditions, i.e., a reflectivity R=1 or under no original conditions, i.e., transmittance R=1. Resultant output signals, which have been subjected photoelectric conversion by the respective cells, are fed by way of the changeover switch 2 to the (+) input terminal of the addition/subtraction circuit 4. Here, the dark output fractions b, of the cells S, assuming corresponding positions are read out successively in synchronization with the camera 1 and are then D/A converted. By feeding the thus-converted analog signals to the (-) input terminal of the addition/ subtraction circuit 4, the respective dark components b1 are removed from the output signals v; of the corresponding cells Si of the camera 1.

Namely, the following signal is obtained as an output from the addition/subtraction circuit 4: V1-W=C1 R1+b1-b1 =c- R=CX (cf., R= 1 ) The resulting signal is then subjected to logarithmic compression by the logarithmic transformation circuit 7.

By holding the change-over switch 8 turned to the addition/subtraction circuit 9 and expecting the range of variations of the logarithmicallytransformed reference signals log cfi in advance, the constant level E which is smaller than the range of variations is subtracted from the logarithmicallytransformed reference signal log ci at the addition/ subtraction circuit 9. Therefore, the variation Aci (= log c-E) is alone A/D-converted and stored in the memory 12.

For the convenience of explanation, the sensitivity correction value Aci has been obtained under the conditions of R=1. It is also possible to obtain the sensitivity correction value by providing, as an original picture to be input, a reflective or transmissible original the surface of which is uniform in its entirety and is substantially of R=1.

Where a CCD sensor is formed into a onedimensional line sensor, sensitivity correction valuesAC, relating to the two-dimensional illuminance l,i can be successively obtained while shifting the line sensor in the sub-scanning direction.

It is most desirable to perform the first and second steps for each original prior to scanning the same.

These steps may however be skipped if each parameter remains stable.

By the above-described first and second steps, the dark output fractions bi of the individual cells Si and the variations AC, of the white level reference signals are stored respectively in the memory 5 and the memory 12.

These first and second steps make up the abovedescribed preparatory step. The following third step will be continuously performed while the original picture is input.

Step 3: This is a step usually practiced (i.e., operation step of correction of picture input).

When reading in the original picture by the camera 1, the change-over switch 2 is kept turned to the addition/subtraction circuit 4 while the changeover switch 8 is kept turned to the addition/ subtraction circuit 10.

In synchronization with each picture signal vi from the camera 1, the dark output fraction of the cell at its corresponding position is read out from the memory 5. At the addition/subtraction circuit 4, the dark output fraction is subtracted from the picture signal. Thereafter, the logarithmically-transformed picture signal log (Y-Vi) is output. In synchronization with the logarithmicallytransformed picture signal, the variation Aci of the white level reference signal for the corresponding cell position is read out from the memory. The variation AC, is then fed via the D/A converter 13 to the addition/subtraction circuit 10, where it is subtracted from the corresponding logarithmicallytransformed picture signal.

The above-described operation process may be expressed by the following equation: log (V-V1)-AC1-E =log (Ci R1+b,-b1)-(log q-E)-E =log C Ri-log ci =log c,+log R,-log q =log Ri (6) In the above manner, variations of the dark output fractions of the individual cells Si and the convolution values cj of the respective parameters, which convolution values correspond to overall gamma characteristic values, are removed. Hence, signals depending on the reflectivity or transmittance Ri of the original are obtained as the outputs log R1 of the addition/subtraction circuit 10.

In the above embodiment, the picture signals input in Step 3 are processed as they are, i.e., as analog signals. Thus, the outputs log Ri are obtained as analog signals of the picture density signals per se. If the thus-input picture signals are subjected to further analog processing by subsequent units, the reliability will become extremely high.

Reference is next made to Figure 2 which illustrates the second embodiment of this invention.

In the second embodiment, the second arithmetic circuit, i.e., the addition/subtraction circuit 14 has been digitized. An A/D converter 15 is provided between the logarithmic transformation circuit 7 and the change-over switch 8, whereby omitting the D/A converter 13 employed in the first embodiment.

By constructing a picture input equipment in the above-described manner, the second embodiment of this invention has brought about such merits that the D/A converter 13 has been omitted, the outputs log Ri can be obtained as digital signals, and where subsequent processing units are digital equipment, a suitable picture input equipment can be obtained.

Referring next to Figure 3, the third embodiment of this invention will be described. In the third embodiment, each output signal which has been obtained owing to photoelectric conversion by the camera 1 is converted immediately to a digital signal by means of an A/D converter 15 so as to omit the D/A converter 6 employed in the first and second embodiments.

Accordingly, the signal processing subsequent to an addition/subtraction circuit 4' are all performed in digital fashion. A logarithmic transformation circuit 16 is also composed of a digital logarithmic transformation table.

By constructing a picture input equipment in the above-described manner it is possible to omit the ND converter and D/A converter which are provided respectively before the memory 5 and after the memory 12. Thus, the circuit structure can be simplified further.

The present invention may be practised by making various changes or modifications, besides the above-mentioned first to third embodiments.

For example, the memory 5 which is provided to subtract the dark output fractions bi may be omitted by inputting a signal equivalent to the average value of the dark output fractions bi provided that the variations of the dark output fractions bi are ignorably small. Omission of the memory 5 permits further simplification to the circuit structure and from the viewpoint of work steps, permits omission of the first step. Depending on the level of required accuracy, the present invention may be practised in various modified manner.

In each of the above-described embodiments, errors of digitization at the A/D converters 3, 11 and 15, etc. may be reduced if the signals b1 and Acj, which are stored respectively in the memories 5, 12, are stored as their average values by inputting them after scanning the individual cells several times.

Having now fully described the invention, it will be apparent to one of ordinary skill in the art that many changes and modifications can be made thereto without departing from the spirit or scope of the invention as set forth herein.

Claims

1. A picture input equipment comprising: a CCD array sensor adapted to perform photoelectric conversion based on the quantities of incident light to respective cell; a first arithmetic circuit for subtracting dark output fractions from corresponding signals obtained from the array sensor; a logarithmic transformation circuit for transforming signals output from the first arithmetic circuit into logarithmic values; memory means for storing logarithmicallytransformed reference signals on the basis of signals output respectively from the cells when incident standard light of white level is input to the array sensor; and a second arithmetic circuit for correcting picture signals input from the cells of the array sensor and subjected to logarithmic transformation by either adding or subtracting white level reference signals, which are read out respectively for the corresponding cells from the memory means, to or from the picture signals.

2. A picture input equipment comprising: a CCD array sensor adapted to perform photoelectric conversion based on the quantities of incident light to respective cell; first memory means for storing dark output fractions corresponding respectively to cells of the array sensor; a first arithmetic circuit for subtracting the dark output fractions, which are stored in the first memory means from signals output from the array sensor; a logarithmic transformation circuit for transforming signals output from the first arithmetic circuit into logarithmic values; second memory means for storing logarithmically-transformed reference signals on the basis of signals output respectively from the cells when incident standard light of white level is input to the array sensor; and a second arithmetic circuit for correcting picture signals input from the cells of the array sensor and subjected to logarithmic transformation by either adding or subtracting white level reference signals, which are read out respectively for the corresponding cells from the second memory means, to or from the picture signals.

3. Picture input equipment substantially as herein described and shown in the accompanying drawings.