JPH07240817A - Method for detecting defective element of image sensor and image signal correcting method - Google Patents

Abstract

PURPOSE:To correct an image signal read by a CCD sensor with high precision by detecting a defective element of the CCD sensor whose detection has been difficult in the stage of the manufacture and shipment. CONSTITUTION:A gradation original 4 is set to an original platen 2 of a plane input scanner 1 and scanned in the subscanning direction. A transmission light 14 whose incident light quantity is changed continuously is made incident in a CCD line sensor 10 being a check object. An image given by an output signal v1 of each element is printed on a film 32 by an output scanner 31. A defective element is detected by recognizing a stripe image in the image of the film 32. The image of position information representing the arranged position of each element is printed on the film 32 and both image on the film 32 are compared to specify the arranged position of the defective element. Then the image signal of the defective element is corrected by synthesizing both the image signals on the defective element and elements in the vicinity of the defective element.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for detecting a defective element of an image sensor and a technique for correcting an image signal read by using an image sensor having such a defective element. In particular, the invention is
For example, the present invention relates to a technique suitable for a field in which it is required to accurately read an image of a document such as a field of photolithography. Further, the present invention also relates to the field of an inspection apparatus that detects a defective element included in the image sensor with high accuracy in advance so that the image sensor can be applied to the field of photolithography.

[0002]

2. Description of the Related Art In an image recording apparatus such as a facsimile, a copying machine, an electronic blackboard, and a plate-making scanner, a CCD (Charge Coupled) is used as an image sensor for reading an original image.
Device) line sensor and CCD area sensor (hereafter
Both are collectively referred to as CCD sensors) are widely used. However, in the field of plate making scanners and the like, since a high quality recorded image is required, a CCD sensor including a defective element showing an abnormal output characteristic cannot be used as an image sensor of the plate making scanner. Therefore, a CCD sensor is inspected in advance by itself to select a problematic CCD sensor.

As a document disclosing such a method of inspecting a CCD sensor, there is Japanese Patent Laid-Open No. 60-197064. Moreover, the document also discloses a method of correcting the output value of the defective element. First, this document is the second
As described in Table 1 in the lower left column of the page and in line 4 to the lower right column of the same page to the eleventh line of the same column, an element that gives a large or small output value over a certain range with respect to incident light of a constant intensity. Are included in the CCD sensor, but the number thereof is small, and one of the causes of causing an abnormality in such an output value is a short circuit of the electrode of the element or a disconnection of the wiring. It discloses that you can do it. Then, as a method of detecting such a defective element and correcting the output value thereof, the document proposes the following method. That is, page 2, upper left column, line 15 to same page, upper right column, third line
Line and the upper left column, line 15 of page 3 to the upper right column, line 17 of the same page, each element of the CCD sensor is irradiated with light of constant intensity, and the output value of each element and the average of the output values of all elements are averaged. The presence or absence of a defective element is detected based on whether the difference from the value is within X% of the average value. When it is within X%, it is determined that there is no defect, and the output value is multiplied by the correction factor γ for correction.
The correction by the correction factor γ corresponds to the shading correction generally performed in image processing. On the other hand, when it does not fall within X%, it is judged as a defective element, the number of the defective element is stored, and the output value of the element immediately before the defective element or the average of the output values of the elements before and after the defective element is stored. The value is used as the true output value of the defective element.

This prior art is not suitable for a CCD in advance in a device such as a plate-making scanner which is required to reproduce a high quality image (for example, a device which needs to read an image with a resolution of 12 bits). It can be used as a detection method for detecting a sensor and excluding such a CCD sensor from a product.

[0005]

According to the above conventional technique,
It is possible to detect a considerable number of improper CCD sensors in advance at the stage of manufacturing and shipping the device. However, even if a CCD sensor determined to have no defective element by the above-mentioned conventional technique is used as an image sensor for a plate making scanner, white scratches and blacks are still present in an image actually printed on a sensitive material such as film. An image with streaky scratches called scratches or the like is generated along the sub-scanning direction in which the CCD sensor should be scanned. Such streak-like scratches significantly deteriorate the image quality. The fact that such a phenomenon actually occurs frequently confirms that an undetected defective element still exists in the CCD sensor even with the above-described conventional technique. .

Here, in the case of using the above-mentioned prior art, that is, in the method of detecting the presence or absence of a defective element only from the output value of the CCD sensor alone, the detectable resolution is the information amount of the image signal. In the case of seeing,
It is considered that 8 bits (256 gradations) are the limit. Therefore, the value of the parameter X (%), which is a reference for determining the presence / absence of a defect in the above conventional technique, is about 10% at most. On the other hand, as an example of the resolution required for an image recording device such as a plate-making scanner, for example, a plane scanning device has about 12 bits. Therefore, in a plate making scanner or the like, the parameter X (%) is about 0.0
It means that the value of 25% is required. In other words, in order to detect the defective element that causes the above-described streak-like scratches, it is necessary to make a judgment with a high accuracy of about 0.025%.

Such a minute change in the output value due to a defective element is usually at a level of being buried in electrical random noise (so-called 1 / f noise, reset noise, etc.), and each CCD sensor It can be said that it is difficult to detect by the method of directly measuring the output value of the element. However, even with such a minute level of defective element, when a CCD sensor including the defective element is applied as an image sensor to a high quality image recording apparatus such as a plate making scanner, it can be clearly recognized by human eyes. Unnecessary scratches appear on the sensitive material.

As described above, the present invention exceeds the limit of the inspection accuracy at the stage of manufacturing / inspecting the image sensor,
Therefore, the present invention provides a method for surely detecting the presence / absence of defective elements at a level that cannot be sorted and recognized at the shipping stage. Another object of the present invention is to provide a detection method that enables not only the presence or absence of a defective element but also the position of the defective element to be specified. Further, the present invention provides a correction method for correcting the output result of the defective element with high accuracy based on the position information of the defective element so as to record a high-quality image without streaky scratches. Doing is one of the purposes.

[0009]

According to a first aspect of the present invention, light is incident on an image sensor having a plurality of elements that perform photoelectric conversion while changing the amount of incident light and the output signals of the plurality of elements are measured. At the same time, an image given by each output signal is recorded on a sensitive material, and based on the recorded image, it is detected whether or not the image sensor has a defective element that causes an abnormality in the output characteristic with respect to the incident light amount. There is.

According to a second aspect of the invention, at the time of recording the image according to the first aspect of the invention, in addition to the images of the output signals of the plurality of elements, the arrangement position of each of the elements in the image sensor is shown. An image of position information is also recorded on the photosensitive material, and by comparing both recorded images of the position information and the output signal, in addition to detection of the presence or absence of the defective element, the array position of the defective element is detected. We are also identifying.

According to a third aspect of the present invention, (a) an image sensor having a plurality of elements for performing photoelectric conversion, injecting a constant amount of light, and (b) abnormalities in a region having an output characteristic with respect to the incident light amount. Is detected based on the output signals of the plurality of elements, and (c) the amount of incident light in the step (a) is detected. A step of changing and using light having a new light amount, repeating both the steps (a) and (b).

According to a fourth aspect of the invention, the step (b) in the third aspect of the invention is (b-1) where the level of each of the output signals is within an allowable range determined according to the incident light amount. And (b-2) when the level is not within the allowable range, it is determined that the corresponding element is the defective element.

According to a fifth aspect of the present invention, in the step (b) in the third or fourth aspect of the invention, when the element to be detected is the defective element,
The method further includes the step of identifying the arrangement position of the defective element in the image sensor and recording the position information giving the arrangement position.

According to a sixth aspect of the present invention, in the image signal giving the image of the original read by the image sensor,
An image signal correction method for correcting an image signal corresponding to a defective element that causes an abnormality in the output characteristic with respect to the amount of incident light among a plurality of elements that perform photoelectric conversion included in the image sensor, and corresponds to the defective element. The image signal and the image signal corresponding to the element located near the defective element are combined, and the image signal obtained by the combination is determined as a new image signal corresponding to the defective element.

[0015]

[Action]

(Invention of Claim 1) When light is incident on the image sensor while changing the amount of incident light, each element in the image sensor normally outputs an output signal according to the amount of incident light. However, in the case of a defective element, its output signal does not show a value according to the amount of incident light and becomes an abnormal value. When an image provided by such an output signal is recorded on a light-sensitive material, an image caused by an abnormality in the output signal of the defective element emerges in the recorded image, and it is in a state where it can be visually recorded.

(Invention of Claim 2) When the image of the position information of each element is recorded on the sensitive material together with the image of each output signal, even if the output signal of the defective element is abnormal in the recorded image. Since the fluttered image emerges and emerges, the contrast relationship between the image by the defective element and the array position of the element given by the image of the position information becomes visually recognizable.

(Invention of Claim 3) When a certain amount of light is incident on the image sensor, each element usually outputs an output signal corresponding to the amount of incident light. However, with respect to the defective element, the output signal of the defective element will be abnormal for a certain amount of incident light. Therefore, if there is no abnormality in the input signal of the element for the incident light amount, it is detected that the element is not a defective element, and conversely,
If the element is abnormal, the element is detected as a defective element. Therefore, by changing the amount of incident light, whether or not a defective element exists is detected for each amount of incident light based on the output signal of each element.

(Invention of Claim 4) If the level of the output signal of each element is within the allowable range determined by the incident light amount, the output signal is a normal value according to the incident light amount. , The element is not a defective element. However, when the level of the output signal is out of the allowable range, it can be determined that the output characteristic of the element shows an abnormality with respect to the incident light amount, and the element is detected as a defective element. Therefore, by changing the amount of incident light and determining for each incident light amount whether or not each output signal is within the allowable range at that time,
The presence of all defective elements is detected.

(Invention of Claim 5) Since it is possible to detect whether or not the element is a defective element based on the output signal of each element, when the defective element is detected, the array of the defective elements in the image sensor is detected. The position can also be specified. Therefore, the position information giving the array position is performed subsequent to the detection of the presence or absence of the defective element.

(Invention of Claim 6) An image signal corresponding to an element of each defect is an image signal obtained by synthesizing the image signal and an image signal corresponding to an element located in the vicinity of the defective element. Will be replaced.

[0021]

【Example】

[Points of Interest] First, the causes of the phenomena of white scratches and black scratches described in the section “Problems to be solved by the invention” are analyzed, and the points of focus of this embodiment will be described.
In addition, the white flaw and the black flaw are output voltage characteristics (hereinafter, referred to as "output voltage characteristics" of some elements (a portion that performs photoelectric conversion) in the CCD sensor.
Output characteristic) shows an abnormal behavior, and therefore, it appears that an image corresponding to the defective element appears to have streak-like scratches on the image printed on the photosensitive material.

First, FIG. 1 is a diagram showing an example of the output characteristics of a certain element having a normal output characteristic with respect to the amount of incident light, and an output voltage proportional to the amount of incident light is obtained. The output voltage saturates at the saturated output V _S when the amount of incident light exceeds a certain level. Therefore, in general, each element of the CCD sensor has a variation in sensitivity (a phenomenon that the inclination of the solid line SL0 shown in FIG. 1 is different for each element), and the light receiving surface of the CCD sensor is irradiated. Since the amount of the emitted light is not uniform on the light receiving surface due to the light amount distribution of the light emitted from the light source,
As shown in (a), the output characteristics of each element do not have a certain fixed characteristic. In FIG. 2A, the output characteristics L1, L2, and L3 of the three elements show linearity with respect to the amount of incident light, but the inclinations of the characteristics are different. As shown in FIG. 2B, the problem of such variations in the output characteristics is as follows. As shown in FIG. 2B, the shading correction is performed by setting two points of the white reference (highlight) and the black reference (shadow), and the linearity of each element It is solved by normalizing the output characteristics. In FIG. 2B, the relationship between the output voltage S ′ after shading correction and the amount of incident light is normalized to the output characteristic represented by the straight line L0. The points described above are true for elements having normal output characteristics.

In view of the above-mentioned points, the applicant of the present application considers that the cause of the white scratches and the black scratches is that a certain element exhibits a non-linear output characteristic within a limited range of the incident light amount. Then, it is considered that the non-linear fluctuation of the output voltage is such a minute level that it is usually buried in the random noise (reset noise, 1 / f noise, etc.) component included in the output voltage of the CCD sensor. This point is shown in FIGS. 3 and 5.

First, FIG. 3 shows the output characteristics of a certain element which is considered to cause white defects when a positive film is used as the photosensitive material. The applicant of the present application believes that when a negative film is used as the photosensitive material, the output characteristics shown in FIG. 3 cause black defects. In the figure,
The normal output characteristic (linearity) that should be present is shown by a broken line BL (a part of which is superposed on the solid line SL1). On the other hand, the output characteristic of the defective element that causes a white defect is shown by the solid line SL1, and the output voltage thereof is larger than the normal output characteristic BL at a certain incident light amount. It is considered that this non-linear portion causes a white defect. Then, in this non-linear portion, the voltage difference between the output voltage and the output voltage in the normal case is inferred from the occurrence of white scratches during printing and the level of measurement accuracy or resolution in the conventional technique. , Saturation output V _S 0.1%
It is considered to correspond to the degree.

FIG. 4 shows an example of the case where the output characteristics of the defective element shown in FIG. 3 are shading-corrected like the output characteristics of other normal elements. In FIG. 4A, two of the three elements have normal output characteristics L4 and L5, and the remaining one element is considered to be a defective element, and its output characteristic shows a non-linear characteristic L6. I assume that. When shading correction is performed on these output characteristics L4 to L6, FIG.
As shown in (b), the normal output characteristics L4 and L5 are similarly normalized to the linear output characteristic L0, while the non-linear output characteristic L6 shows the non-linear output characteristic L01 after the shading correction. Will remain. Therefore, it is considered that the non-linear portion AR1 in the output characteristic L01 causes white defects during printing.

On the other hand, FIG. 5 shows an example SL2 of the output characteristics of a defective element, which is considered to be the cause of black scratches (white scratches in the case of a negative film) that occur when the output pixels of the CCD sensor are printed on a positive film. FIG. 6 is a diagram drawn together with the output characteristic BL in a normal case. In the case of a black defect, a certain amount of incident light (in particular, the amount of incident light itself is the saturated output V _S
When the light intensity is smaller than the incident light intensity when the light is applied), the output voltage is significantly reduced compared to the normal state, and the normal output voltage cannot be output unless it exceeds a certain fixed light intensity.
It is considered that this is due to the localized non-linearity of the output characteristics. Therefore, it is considered that the output voltage of the defective element in this non-linear region is also buried in the random noise component and cannot be directly detected.

Even if the non-linear output characteristic which is considered to be the cause of such a black defect is shading-corrected, it is clear that the corrected output characteristic is also non-linear.
An example of this point is shown in FIG. The same figure (a)
Is the output characteristic L in which two of the three elements are normal
7 and L8, and the other one is an example showing a non-linear output characteristic L9. As a result of the shading correction, as shown in FIG. 7B, the output characteristics L7 and L8 are normalized to the linear output characteristic L0, and the output characteristic of the defective element remains the non-linear output characteristic L02.

Regarding the number of elements having nonlinear output characteristics as analyzed above, one CCD sensor
If the total number of elements is 5000, for example, it is estimated that several elements have such a defect.
Therefore, when a document image is read by a CCD sensor including such a defective element, a streak appears in the read image of the portion corresponding to the defective element.

This point is schematically shown in FIG. 7 as an example of printing on the positive film 91. The figure (a) is CC
It is shown that the D sensor 100 includes a white flaw element E1 which is considered to give a white flaw and a black flaw element E2 which is thought to give a black flaw, and FIG. The image of 90 is shown, and the figure (c) shows the printing result. White scratches, black scratches, that is, white stripes and black stripes (92, 9 in FIG. 7C)
3) is not completely white or black, but as a result of comparison with the density of the image around it, it is visually "whiteish",
It is judged as "blackish" and looks like a streak.

As described above, if attention is focused on the fact that the phenomenon of white flaws and black flaws is caused by the minute non-linear output characteristics of the element, such non-linear output characteristics of the defective element are positively considered. The presence or absence of the defective element and the C of the defective element can be visually revealed.
It is considered that the specific position of the array in the CD sensor (hereinafter, simply referred to as the position of the defective element) can be detected. One of them is to use an output scanner to emphasize the shift between the non-linear characteristic and the linear characteristic with high accuracy on the film, thereby visualizing the non-linearity. Secondly, the accuracy of reading by the scanner is increased to remove random noise components, and the buried non-linear output characteristics are electrically revealed. In either method, instead of inspecting the CCD sensor as a single unit, the CCD sensor is installed in the scanner and combined with the function of the scanner body (including the output scanner),
It can be said that there is a feature in trying to detect a defective element in the CCD sensor. Hereinafter, details of the two detection methods will be sequentially described.

[First Detection Method] This method is classified into a method of visually determining the presence or absence of a defective element and a method of visually identifying the position of the defective element. Therefore, a so-called stand-alone type scanner in which an input scanner for reading an image and an output scanner for printing the read image on a film are integrally configured is not used as an image recording device but is included in a CCD sensor. Used as a detection device for defective elements. This is based on the viewpoint of actively applying the original image recording function of the stand-alone type scanner to the detection function. The scanner is applied to any of the above detection methods. In the following embodiment, for convenience, a CCD sensor as an inspection object is
Only the CD line sensor will be described. Of course, this detection method can also be applied to a two-dimensional CCD area sensor.

First, FIG. 8 is a block diagram schematically showing the configuration of the above-described inspection device for defective elements of the CCD line sensor, and the flat input scanner 1 is used as the input scanner. This inspection device is roughly divided into
The flat-type input scanner 1 includes an image processing unit 21, a color conversion processing unit 30, an output scanner 31, and a control unit 13. Hereinafter, the configuration of each unit will be sequentially described.

The flatbed input scanner 1 includes a document table 2, a light source section 24, a drive source section 25, an XY table 9 and an image reading section 8.
And mirror 7. The light source unit 24 includes a halogen lamp 5 and a rod illuminator 6, and the halogen lamp 5 is driven by the control unit 13 described above. A mirror 7 is arranged directly below the rod illuminator 6 via the document table 2. An opening is formed in the center of the document table 2, and a transparent glass plate 3 is fitted in the opening. Then, the gradation original 4 is placed on the glass plate 3. The document table 2 can be moved in the sub-scanning direction X by a drive mechanism (not shown). That is, the movement of the document table 2 causes the gradation document 4 to be scanned in the sub-scanning direction X.

On the other hand, a zoom lens 13 and a CCD line sensor 10 to be inspected are mounted on the XY table 9. The XY table 9 can be moved in the main scanning direction Y by a drive mechanism (not shown) for trimming adjustment. The zoom lens 13 is arranged so as to face the mirror 7, and transmits the transmitted light 14 (hereinafter also referred to as incident light 14) transmitted through the gradation original 4 and reflected by the mirror 7 to the light receiving surface of the CCD line sensor 10. That is, an image is formed on each element. Further, the zoom lens 13 has a diaphragm (not shown). The image reading unit 8 uses the above-mentioned C
CCD line sensor 1 including CD line sensor 10
0 drive circuit 11 and A / D converter 1 for A / D converting the output voltage of each element of the CCD line sensor 10.
It consists of two. The drive circuit 11 is controlled by the CPU 14 described later.

On the other hand, the drive source section 25 is composed of five pulse motors PM1 to PM5. Each pulse motor PM1-P
M5 receives the drive signal output from the control unit 13, and outputs the drive force to the corresponding units. That is, the pulse motor PM1 outputs the driving force to the drive mechanism of the document table 2 described above to move the document table 2 in the sub-scanning direction X.
The pulse motor PM2 outputs its driving force to the zoom lens 13
Output to and adjust the magnification. Also, the pulse motor P
The M3 outputs the driving force to the driving mechanism of the XY table 9 described above to perform trimming adjustment. The pulse motor PM4 outputs the driving force to the zoom lens 13 to adjust the focus position. Also, the pulse motor PM5
Outputs the output to the diaphragm of the zoom lens 13 to adjust the diaphragm.

The image processing section 21 receives the A / D converted image signal V _1, and performs shading correction, logarithmic conversion (LOG conversion), filtering processing by a noise filter, and sub-scanning position on the signal V ₁ . This is a part that sequentially performs each process such as correction, and the timing control unit 22 controls these processes. Then, the timing control unit 22
The control unit 13 controls itself. still,
The memory 23 is for storing a signal V _PI (positional information signal) indicating the positional information of each element or the information for giving the numbering, and is used only in the case of the above detection method.

The color conversion processing unit (CU unit) 30 subjects the image signal V ₂ output by the image processing unit 23 to contour enhancement processing or the like on the signal V ₂ , and then Y (yellow), M This is a portion that is converted into an image signal V ₃ that gives each print color of (magenta), C (cyan), and K (black). Since the operation of the processing unit 30 is well known, its details are omitted here.
In this embodiment, the CCD line sensor 10 to be inspected
Is not a so-called color CCD sensor, but if it is a color CCD sensor, the output signals of each color of B (blue), G (green), and R (red) output from the color CCD sensor are output to the printing colors of YMCK. Output from the image processing unit 21 without being converted into the signal of the scanner 3
Set to output to 1.

The output scanner 31 modulates the light beam based on the image signal V ₃ and then applies the light beam to the film 3
2 is a device (image recording device) for printing the image of the gradation original 4 read by the CCD sensor 10 on the film 32 by scanning and exposing the image on the film 2. Therefore, the output scanner 31 converts the image signal V ₃ into a halftone dot signal of a binary level, for example, a so-called dot generator, a light emitting element (LD, LED, etc.) driven according to the halftone dot signal, or light emission. It has an optical system for forming an image of the light beam emitted from the element on the film 32, a cylinder that rotates in the main scanning direction, and has the film 32 mounted on the outer surface thereof. The light emitting element and the optical system are provided in an exposure head that is movable in the sub scanning direction.
The structure of such an output scanner is well known, and a detailed description thereof will be omitted here.

The control section 13 is a section for controlling the above-mentioned plane type input scanner 1, image processing section 21, color conversion processing section 30 and output scanner 31, and the CPU 14 is the core thereof. In addition, control unit 1
Reference numeral 3 denotes a motor control unit 15 and a motor driver 16, which generate a drive signal to be output to the drive source unit 25, and a memory 1.
7 has an I / F circuit 18. Further, an input device 20 such as a keyboard and a mouse and a monitor 19 such as a CRT display device are connected to the CPU 14 via the I / F circuit 18.

Next, the procedure of the above detection method will be described with reference to FIG. 8 and the flow charts shown in FIGS. 9 and 10.

(Step S1) Prepare a gradation original 4 whose density gradually changes from shadow density to highlight density.

(Step S2) The CCD line sensor 10 to be inspected is arranged on the XY table 9 of the flatbed input scanner 1 via a socket (not shown).

(Step S3) The glass plate 3 of the document table 2
The gradation original 4 is placed on the top of the gradation original 4 so that the original density gradually changes in the sub-scanning direction X.

(Step S4) After adjusting the diaphragm and magnification of the zoom lens 13 and trimming the XY table 9, the image of the gradation original 4 is read next. That is, the halogen lamp 5 is turned on to irradiate the gradation original 4 with light emitted from the rod illuminator 6, and the original 2 is moved in the sub-scanning direction X to scan the gradation original 4. As a result, the transmitted light 14 whose amount of light continuously changes in accordance with the scanning is the CCD.
It is incident on each element of the line sensor 10.

(Step S5) After the image signal V ₁ is subjected to image processing such as shading correction and color conversion processing, the image of the gradation original 4 is printed on the film 32 based on the image signal V ₃ , and then developed. .

(Steps S6 to S8)
This is a visual judgment step of the operator. The operator looks at the printing result (recorded image) of the film 32 output from the output scanner 31 and determines whether or not there are streak-like white scratches or black scratches in the image described on the film 32. . If those scratches are observed, the operator is the CCD line sensor 1 concerned.
On the other hand, it is determined that 0 has a defective element, and when those scratches are not observed, it is determined that there is no defective element. The operator may input the judgment results to the control unit 13 using the input device 20.

(Step S9) The operator judges whether or not all the CCD line sensors 10 to be inspected have been inspected. If all the CCD line sensors 10 have not been inspected, then the steps S2 to S8 for the next new CCD line sensor 10 are performed. repeat. Then, when all the CCD line sensors 10 have been inspected, the present detection method ends.

Next, the detection method will be described. The determination of the presence or absence of a defective element in this detection method is the same as that of the detection method, but in this detection method, it is also possible to specify the position of the defective element by printing the position information of each element on the film. There is. This point will be described in detail first.

FIG. 11 and FIG. 12 are images giving the position information for specifying the position of each element together with the read image of the gradation original 4 (FIG. 8) by using the output scanner 31 shown in FIG. Also schematically shows the result (recorded image) printed on the film 32. Of these, FIG. 11 is a diagram showing the entire film 32 after printing and development, in which the image giving the above position information is printed in the first printing area 33, while the gradation original 4 ( The image (halftone image) of FIG. 8) is printed. Here, the CCD sensor is assumed to have N (for example, 5000) elements. Therefore, the number of pixels in the main scanning direction Y read by the CCD line sensor is N. Further, the image giving the above-mentioned position information has a resolution of 13 bits in the direction corresponding to the sub-scanning direction X and has the first printing area 33.
It is burned on. FIG. 12 shows the image giving this position information in more detail, and FIG.
This corresponds to a diagram schematically showing an enlarged area 35 surrounded by a broken line in FIG. 11.

In FIG. 12, each of the images 36 to 39 formed in the first printing area 33 has 1 bit,
The element position information is given with a resolution of 2 bits, 4 bits, and 8 bits. Further, the aggregates 42 with numbers 0 to 11 attached to the upper part of the figure correspond to the numbers of the arrangement positions of the elements, and for convenience of explanation here,
It is written in this drawing. Images like this 36-3
How 9 identifies the position of each element is as follows.

Assuming that a white flaw or a black flaw is present in the direction of the arrow 36 shown in FIG. 12, the extension line of the flaw intersects both the images 37 and 38. Each image 37, 38
Indicates the numbers 2 and 4, respectively, so the value 6 obtained by adding the numbers 2 and 4 shown by both images 37 and 38 corresponds to the number of the array position of the element. Therefore, in this case, N in the CCD line sensor
When the element on one side of the individual elements is standardized as the 0th element, the 6th element counted from the reference element has a defect causing the flaw. The above description is only an example,
Similarly, the defective element can be visually identified. Generally speaking in this respect, the operator is required to use the first printing area 33 that intersects with the extension line of the white flaw or the black flaw.
It is possible to know the position number of the defective element by visually detecting the image inside and adding the numerical values given by the intersecting images.

The position of the defective element can be visually judged by the method shown in FIGS. 11 and 12 in principle, but from the practical application point of view, it is shown in FIG. Further, each image 36, 37, 38, which gives position information with each resolution of 1 bit, 2 bits, 4 bits, ...
The size such as ... Is such that it is difficult for an operator to visually recognize the size accurately. Therefore, in order to overcome this problem, the size of each image printed on the film 32 may be expanded in the main scanning direction Y. Focusing on this point, FIG.
13 and 14 show that the size of each of the recorded images is expanded twice in the main scanning direction Y.

FIG. 13 is an overall view of the film 32 after baking. Here, since the image in FIG. 12 is enlarged or inflated by a factor of two, N / 2 in the main scanning direction Y is obtained.
Since the image can be printed only for each pixel, half of the image giving the position information of the element is first printed in the first printing area 33A, and then the image of the gradation original read by N / 2 elements. Is printed in the second printing area 34A, the image giving the remaining half position information is printed in the third printing area 33B, and the image of the gradation original in the other half is printed in the fourth printing area 34B. FIG. 14 is an enlarged schematic view of the region 35A surrounded by the broken line in FIG. Each image 36 ', 37', 38 ',
, Etc. are doubled, it is easier for the operator to identify the defective element.

It is preferable to set the enlargement ratio or the water increase ratio at the time of baking to about 5 times.

Therefore, based on the above description, the procedure of the detection method will be described below with reference to the block diagram of FIG. 8 and the flowcharts of FIGS.

The gradation original 4 is prepared (step S1A), and the image signal V _PI (hereinafter, referred to as the positional information of each element of the CCD line sensor 10 to be inspected)
(Also referred to as position information signal) is stored in the memory 23 via the input device 20 and the control unit 13. As such an image signal V _PI , the signal which gives the burned image shown in FIG. 12 is used (step S2A). afterwards,
The CCD line sensor 10 to be inspected is arranged in the reading unit 8 (step S3A), the gradation original 4 is set (step S4A), and the gradation original 4 is set.
Image is read and stored in the memory 23 (step S5).
A). After that, the position information signal V _PI and the image signal of the gradation original 4 are read out from the memory 23, both signals are enlarged at a predetermined enlargement rate or a padding rate, and both signals are transmitted to the color conversion processing section 30. Color conversion processing is performed (step S6A), and the image of the position information of each element given by the position information signal V _PI and the image of the gradation original 4 are sequentially printed on the film 32 by the output scanner 31 (step S7A). .

After that, visual judgment is made by the operator. First, it is judged whether or not there are white or black scratches on the film 32 (step S8A), and if there is, it is judged that there is a defective element (step S9A), and the position of the defective element, that is, the above-mentioned position number is determined. Identify. As described above, this is performed by visually detecting an image which gives position information and which intersects with the extension line of the white flaw or the black flaw (step S10A). On the other hand, if there is no flaw, it is determined that there is no defective element (step S11A). After the identification, the operator inputs the detection result of the CCD line sensor 10 (presence or absence of the defective element and the position of the defective element in the presence of the defective element) to the control unit 13 using the input device 20 (step S12A). . In this case, the control unit 13
However, the input result may be displayed on the monitor 19. Then, the above steps S3A to S12A,
This is performed for all CCD line sensors 10 to be inspected (step S13A).

As described above, in the detection method, since the image giving the position information of each element is printed together with the image of the gradation original, the presence or absence of the defective element can be detected without removing the random noise as described above. There is an advantage that the position can be accurately specified. This means that when the random noise is electrically removed, the detection accuracy of the defective element depends on the removal accuracy of the noise by any means. It means that the defective element can be accurately identified without taking any measures to overcome the problem.

[Second Detection Method] In contrast to the former method, which outputs a film and visually checks the presence or absence of a defective element by an operator, in the present method,
The point that the image of the gradation original is read by the input scanner is similar to the previous method, but the following points are different. That is, in this method, the gradation original is scanned with high accuracy to reduce the random noise component from the output signal of each element, and the output signal of each element with an improved S / N ratio, that is, each image signal By electrically performing signal processing, the presence / absence of a defective element and its position are specified. As a result, the method enables accurate detection of defective elements in a relatively short time without printing the image of the original on the film. Therefore, when the present method is applied, the output scanner 31 shown in FIG. 8 is unnecessary and there is no need to use a so-called stand-alone type scanner.

Here, the block diagram of FIG. 18 shows an example of a detection device to which a method of detecting a defective element by such image data processing is applied. 8 is different from FIG. 8 described above in that it has an image processing unit 42 and that there is no output device such as a color conversion processing unit or an output scanner. The other components are the same as the corresponding units in FIG.

First, the control unit 13 (especially its CP
U14) is a part for controlling the scanning of the document table 2 or the gradation document 4 in the sub-scanning direction X. In this method, the control unit 13 controls the scanning of the gradation document 4 as follows. That is, the CCD line sensor 10 scans the gradation document 4 at the same density a plurality of times to read the image in the main scanning direction Y at that position a plurality of times. Stepwise feed in the sub-scanning direction X by a predetermined amount. Then, after performing a plurality of scans in the main scanning direction Y at a new location on the gradation original 4 having the same density, the control unit 13 again feeds the original 4 stepwise. Thus, it is inspected whether or not the output characteristic of the CCD line sensor 10 is abnormal for each density range in the gradation original 4, that is, for each incident light amount of the light 14 incident on the CCD line sensor 10. . The image signals are averaged by a plurality of scans at the same density to reduce random noise.

Further, the image processing section 42 includes an input I / F section 4
3, a shading correction unit 44, a random noise correction unit 45, a detection processing unit 46, and a timing control 49. Of these, the random noise correction unit 45 uses the correlated double sampling (C
DS: Correlated Double Samp
ring) circuit. In addition, the detection processing unit 46
It has a DSP (digital signal processor) 47 and a memory 48. Of these, the memory 48 has four parts
This is a case where the output signals output by the respective elements after the processing by 4, 45 are stored, and the DSP 47
Is (i) the process of averaging the output signals of each element obtained by scanning the same scan line multiple times, and (ii) the average output of the output signals of each element averaged in (i). The process of calculating the value, and (iii)
By comparing the value of the output signal of each element with the average output value, processing for detecting whether each element has a defect or not is performed. Among these processes (i) to (iii),
(i) is a portion for reducing random noise included in the output signal of each element.

Next, the criteria for detecting defective elements in this method will be described with reference to FIG. In the figure, the symbol D _AV represents the average output value of the element at a certain incident light amount. There are two possible methods for calculating the average output value D _AV . One is a method of averaging output signals of all elements, and another is a method of averaging output signals of a plurality of elements adjacent to each other for each arbitrary element to obtain an average output value D. _It is a method of seeking _AV . Here, the description will proceed assuming that the average output value D _AV is obtained by any of the above methods.

Here, if the output signals of the respective elements are printed on a positive film, the levels of the output signals of the defective elements which cause white and black flaws at that time are 50 and 51 shown in FIG. 19, respectively.
become. Therefore, the level D _n of the output signal of any element is compared with the average output value D _AV

[0065]

[Equation 1]

When the relational expression of is satisfied, the element is judged to be a defective element. The symbol α in Formula 1 is a threshold value, and since it is intended for the field of high-precision image recording devices such as photoengraving, a value of about 0.001 is obtained as the threshold value α. To be Therefore, in order to determine that the element is not a defective element, the level D of the output signal of each element is determined.
_n is the next allowable range D _{AV to} D _AV · (1 + α) or D _AV ·
It must be within (1-α) to D _AV . In FIG. 19, since the level D _A giving white defects is increased by α · D _AV or more with respect to the average output value D _AV , the element outputting the level D _A is determined to be a defective element. Similarly, in the figure, the level D _B that gives a black defect is the average output value D.
Since reduced by more than the alpha · D _AV than _AV, this element is also determined as defective elements.

When it is determined that there is a defective element by the relational expression of the equation 1, the DSP 47 determines the number of the element, that is, the position of the element.
It will be detected by a counter (not shown) provided inside. Then, the counter is incremented each time the presence / absence of a defect in each element is determined (determination by the above-mentioned mathematical expression 1).

In FIG. 20, the document table 2 (FIG. 18) is stepwise fed to change the reading position of the image of the gradation document 4 (FIG. 18) in the sub-scanning direction X by a predetermined step amount. FIG. 6 is a diagram showing a state in which a defective element that causes a black flaw or a white flaw is detected when the original density of the original 4 is gradually changed from dark D ₁ (shadow) to bright D ₅ (highlight). . In FIG. 3A, it is assumed that defective elements 52 and 53 that can cause black scratches and white scratches are present in the plurality of elements of the CCD line sensor 10, respectively. Then, when the original density of the gradation original 4 (FIG. 18) is changed stepwise in five steps from the shadow D ₁ to the highlight D _{5 with} respect to the CCD line sensor 10,
That is, the incident light 14 to the CCD line sensor 10 (see FIG.
The incident light amount of 8) is set to the min. Value to its max. The level of the output signal V ₁ of each element obtained when the value is stepwise increased is shown in FIG.
(F). Further, in FIG. 9G, the manner in which the document density of the gradation document 4 (FIG. 18) changes is schematically illustrated in correspondence with FIGS.

First, when the original density is the shadow D ₁ , all of the criteria of the expression 1 are satisfied, and as a result, no defective element occurs (FIG. 20 (a)). However, when the document density is D ₂ (> D ₁ ), it is considered that the output characteristics of the defective element 52 have non-linearity, and therefore the existence of the defective element 52 that causes an abnormality such as black defect occurs.
It is detected according to the criterion of the expression 1 ((b) in the same figure). Further, when the document density is D ₃ (> D ₂ ), both defective elements 5
It is considered that nonlinearity occurs in the output characteristics of Nos. 2 and 53, and the presence of both defective elements 52 is detected by the equation 1 ((c) in the same figure). On the other hand, when the document density is D ₄ (> D ₃ ) in which the document density is further increased, it is considered that only the defect element 53 has non-linearity, and as a result, the defect element 5 that causes an abnormality such as white defect.
The presence of 3 is detected by Equation 1 ((e) in the same figure). Finally, in highlight D ₅ , the output characteristics of all the elements show linearity, and it is determined that there is no defective element (FIG. 6 (f)).

Next, based on the detection principle described above, the procedure of the second detection method will be described in detail based on the block diagram of FIG. 18 and the flowcharts of FIGS. 21 to 23.

(Steps S1B to S3B) These steps S1B to S3B are the same as steps S1 to S3 (FIG. 9) described in the first detection method.

(Step S4B) The CPU 14 receives a scan start command signal (not shown) input by the operator from the input device 20, outputs a drive signal to the pulse motor PM1, and the CCD line sensor 10 reads the drive signal. The original table 2 is moved so that the original document density in the main scanning direction Y becomes a shadow.

(Step S5B) The DSP 47 sends CP
Upon receiving the command signal from U14 and the timing control 49, the count value n of the counter of the DSP 47 is set to 0 (n ← 0). This completes the preparation for scanning.

(Step S6B) The CPU 14 turns on the halogen lamp 5 with a constant intensity to transmit the transmitted light 14
To the respective elements of the CCD line sensor 10. As a result, each image signal in the main scanning direction Y when the document density hits the shadow is converted into the output signal V of each element.
Obtained as ₁ . The output signals V ₁ are sequentially output to the image processing unit 4
2, the shading correction unit 44 receives shading correction, and the random noise correction unit 45 removes random noise such as reset noise (reduces its level), and then the detection processing unit 4
6 is stored in the memory 48. The timing control 49 controls the series of processes.

(Step S7B) Above Step S6B
Step S6B is executed again a plurality of times after the end of step S1 or in parallel. As a result, in the memory 48, a plurality of output signals V ₁ for the document density are output for each element.
Data is stored.

(Step S8B) The DSP 47 takes out the data relating to the corresponding output signal V ₁ from the memory 48 for each element, calculates the average value thereof, and stores the average value again in the memory 48. By this averaging process, random noise is significantly reduced, and it is possible to perform a highly accurate defect element detection process.

(Step S9B) Based on the average value of the output signal V ₁ of each element stored in the memory 48, the DSP 47 averages the average output value D _AV.
Is calculated and stored in the memory 48. Here, the above-mentioned method is applied as the calculation method, and another method will be described later. Therefore, the average output value D _AV is given as a value obtained by dividing the sum of the average values of the output signals V ₁ of all the elements by the number of elements.

(Step S10B) The DSP 47 determines the presence / absence of a defective element for each element in accordance with the criterion of the equation (1). Then, when it is determined that there is no defect (there is within the allowable range), the process proceeds to step S12B.

(Step S11B) When it is determined that there is a defect (outside the allowable range), the DSP 47 determines from the count value n of the counter that it is the position of the element, that is, the nth element counting from the element at one end. The signal that specifies and provides the position information is stored in the memory 4
Store in 8.

(Step S12B) The DSP 47 sends the determination result to the control unit 13. The determination result is information that gives no defect when it is determined that there is no defect, and includes information that gives a defect and information that gives the position of the defective element when it is determined that there is a defect. After that, the CPU 14 stores the determination result in the memory 17
It is stored on the monitor 19 and displayed on the monitor 19. This allows the operator to recognize the presence and position of the defective element.

(Step S13B) The DSP 47 counts up the counter. Therefore, n ← n + 1
Becomes

(Step S14B) The CPU 14 confirms whether or not the read document density is highlight, and if not, the process proceeds to step S15B, and if so, the process proceeds to step S16B.

(Step S15B) Control Unit 1
In step 3, the document table 2 is step-fed by a predetermined amount in the sub-scanning direction X. As a result, the density of the original to be read next, that is, the amount of transmitted light 14 increases. Then, after that, steps S6B to S14B are repeated.

(Steps S16B to S17B) CCD
Remove the line sensor 10 from the image reading unit 8 (S
16B), the above steps S3B to S17B are repeated until all the CCD line sensors 10 have been inspected (S1).
7B).

Next, FIG. 24 showing the positional relationship of each element in the CCD line sensor 10 for the procedure when the above method is applied as the calculation of the average output value D _AV .
And FIG. 25 and FIG. 26 showing a flowchart. However, in this method, the above-mentioned step S9B is performed.
25 to 26, only steps corresponding to steps S9B to S12B are illustrated in FIGS. 25 and 26 because steps other than steps to S12B are the same as the steps shown in FIGS.

As shown in FIG. 24, the number of an arbitrary element is n, the number of elements is m, the level of the output signal of the nth element (processed for shading correction, etc.) is D _n , and the nth element is related. 2A is the total number of elements before and after used to calculate the average output value D _AV.
(A elements are used for each of the front side and the rear side), and the description will proceed.

(Steps S9B1 and S9B2) DSP
47 sets the element number n (which corresponds to the count value of the counter described above) to 0 (S9B
1), n + 1 is replaced with n (S9B2).

(Step S9B3) The DSP 47 determines whether the element number n exceeds the element number m. If n> m, the DSP 47 moves to step S14B (FIG. 23) described above, and otherwise moves to the next step S9B4.

(Steps S9B4 to S9B8) DSP
47, the element number n is n <A, A ≦ n ≦ m−
It is determined whether the relation of A or n> m−A is satisfied (S9B4), and step S9B5 is performed according to each relation.
~ Any one of S9B7 is performed. That is, when n <A, the DSP 47 calculates the average output value D _AV based on the following equation 2.

[0090]

[Equation 2]

When A ≦ n ≦ m−A, the DSP 4
7 calculates the average output value D _AV based on the following equation 3.

[0092]

[Equation 3]

When n> m−A, the DSP 47
Calculates the average output value D _AV based on the following equation 4.

[0094]

[Equation 4]

After the calculation, the DSP 47 stores the average output value D _AV in the memory 48 (S9B8).

(Steps S10B1 to S12B1) D
The SP 47 determines whether or not the n-th relevant element is a defective element based on Equation 1 (S10B).
1). When it is determined that there is, the DSP 47
Stores the element number n in the memory 48 (S11B1). Further, thereafter, the DSP 47 outputs the determination result (information indicating that there is no, information indicating that and the number n if there is), to the control unit 13.
Then, as a result, the control unit 13 displays the above determination result on the monitor 19 and makes the operator recognize the determination result (S12B1).

Applying the method described above is advantageous in the following cases. That is, if each density of the prepared gradation original is uniform with respect to the main scanning direction Y, there is no problem. Will be buried. At this time, in order to eliminate the influence of this nonuniformity, it is desirable to obtain the average output value D _AV for each element by using the output signals of the adjacent 2A elements in accordance with the detection method of. .

In the above method, when the level of the output signal of the defective element is abnormally high or low, the abnormal value is likely to affect the calculation of the average output value D _AV. Can overcome the above problem and increase the detection accuracy by increasing the total number of adjacent elements 2A used for averaging as much as possible.

(Modifications of First and Second Detection Methods) In both of the first and second detection methods described above, the gradation original is scanned to change the light amount of the incident light of the CCD line sensor. However, instead of this method, the following method may be used. That is, a document of uniform density is prepared in advance and set on the document table, and this time the document table is fixed without moving in the sub-scanning direction.
Instead, the halogen lamp 5 in FIG. 8 and FIG.
The amount of incident light of the CCD line sensor is made variable by adjusting the amount of light or the aperture of the zoom lens 13. Also in this case, the output signal V of each element
By performing the same processing as that described in the first and second detection methods on ₁ (FIGS. 8 and 18), the presence or absence of a defect in each element and the position of the defective element can be specified.

As a reference, the procedure when the above-described modification method is applied to the second detection method of electrically processing is shown in FIG.
However, the reference numerals shown in the figure are the same as those in FIG.

Instead of using a gradation original or an original of uniform density, light emitted from a light source such as a halogen lamp may be directly incident on each element of the CCD line sensor through a diaphragm. In this case, the amount of incident light can be changed continuously or stepwise by changing the current applied to the halogen lamp and the diameter of the diaphragm.

[Image Signal Correction Method] By using the method of the first detection method or the second detection method described above, the position of the defective element included in the CCD line sensor can be specified with high accuracy. It has become possible.

Therefore, even if the image signal of the original document to be recorded read by the CCD line sensor is recorded on a sensitive material such as a film, a defect in the CCD line sensor will not occur so that no black or white scratches will occur. It is required to appropriately correct the image signal corresponding to the element based on the position information of the defective element obtained by the above detection method. Moreover, the correction that simply prevents the occurrence of black and white scratches is not sufficient, and as this correction method is applied to an image recording device such as a scanner, the image recorded by the corrected image signal is the surrounding image. Therefore, there is a demand for a correction method capable of reproducing a high quality image in harmony with the above. Below, the procedure of a suitable correction method will be described after examining the principle of a suitable correction method capable of satisfying such requirements.

Although various correction methods can be considered, the following two simple correction methods can be given.

One of them is that the level value of the output signal of the element located one before the defective element is
It is used for the level value of the output signal of the defective element. FIG. 28 schematically shows this method as an aid in understanding this point.

FIG. 11A shows that the nth element in the CCD line sensor 10 is a defective element, and FIG. 11B shows the (n-1) th element before correction. It is shown that the levels of the output signals of the nth and (n + 1) th elements are D _n-1 , D _n , and D _{n + 1} , respectively. In this method, the level D _{n is changed} to the level D
_Since the replacement process is performed with _n-1 , the result after the correction is shown in FIG. In this case, the level value D _n-1 of the element immediately before the defective element continues for two pixels.

The other method is to use the average value of the output signal levels of both elements before and after the defective element as the level value of the output signal of the defective element. FIG. 29 schematically shows this point as well.

28 (a) and 28 (b) are respectively shown in FIG.
This corresponds to (a) and (b). And FIG. 29 (c)
Indicates the level value of the output signal of each element after correction. That is, the level value D _n 'of the new output signal of the n-th defective element is given by the average value (D _n-1 + D _{n + 1} ) × 2.

Each of the above methods is based on the idea of correcting the output signal of the defective element with the output signals of the elements located before and after the defective element. It is similar to the method disclosed in Japanese Patent Publication No. 197064. Certainly, it can be said that the correction method based on such an idea is a simple method and a practical method. However, when the above correction method is applied to an apparatus that records an image with high precision such as a plate-making scanner, a new problem arises in that a high-quality image cannot be obtained. This is because the output signal of the element immediately before the defective element or the average value of the output signals of the elements before and after the defective element is used as a new output signal of the defective element. For example,
D n _{+ 1} >> D when _n-1 magnitude relationship like that is established, the level D when replacing _n-1 to a new level D _n, a new level D _n and the next level D _{n +} The difference from ₁ is too large, and the average value of D _n-1 and D _{n + 1} is set to the new level D.
Even _n, the difference between the (D _n -D _n-1) and the difference (D _{n + 1} -D n) becomes a relatively large value, as a result, in the portion of the defective elements, corrected That is, the difference in the image density, that is, the edge portion is visually conspicuous.

FIG. 30 shows this point directly. This figure shows the result of scanning an oblique line when the original image is an oblique line, correcting the image using the above-described correction method or method, and printing the corrected image signal on a film or the like. . The figure (a) shows that the n-th element is a defective element, and the figures (b) and (c) show the results when the method and are applied, respectively. In the same figure (b), that is, in the case of the method, white and black are interchanged at the corrected portion, so the edge of the corrected portion is conspicuous. Also, in the case of FIG. 6C, that is, in the case of the method, the corrected portion becomes gray, so that the edge of the corrected portion is considerably conspicuous even if it is not as large as in FIG. .

The cause of such a problem is
I think that it is none other than the correction method and since the following points are not considered at all. As described above, the applicant of the present application is
It was pointed out that the point that nonlinearity occurs in the light quantity region of the output characteristics of the element with respect to the incident light quantity is the cause of the white and black scratches, that is, the mechanism of occurrence. For example, it is as shown in FIGS. As can be seen from these figures, the elements judged as defective elements are
It does not always give non-linearity. On the contrary, the nonlinear region is localized in a limited region when viewed from the total incident light amount region. In other words, even if the element is determined to be a defective element, a normal output signal may be output during scanning of the document. This means that, like the correction method or the like, a correction method such that the output signal of the defective element is replaced with the output signal of the element around the defective element at all times is not always required. It can be said that it does not correspond to a suitable correction method. On the contrary, it is necessary to correct the output signal of the self by using the output signal of the defective element itself. Therefore, the above correction method,
Can be said to have not paid any attention to such points.

The applicant of the present application intends to change the way of thinking in view of such points of view, and proposes a correction method capable of effectively utilizing the information provided by the defective element. That is, using both the output signal of the defective element itself and the output signals of the surrounding elements (which means that they are combined at a certain combining ratio), correction is made according to the criteria described below. Is going.

The following correction method is a method that corrects the output signal of the defective element with high accuracy and does not affect the image quality of the edge portion of the recorded image. This method

[0114]

[Equation 5]

It is carried out based on Thus, depending on the value of the parameter β, the level D _n of the output signal of the defective element is used as it is as the corrected level D _n ″, or the level D of the output signal of the element located before or after the defective element is used. It is determined whether _n-1 or D _{n + 1} is used as the corrected level D _n ″. It will be shown below that this correction method can be used for correction so as to obtain higher quality and higher quality image quality than the above correction method.

Therefore, as the first comparative example for that,
Consider the case of a picture with contrast as shown in FIG. In this case, the relationship of D _{n + 1} > D _n >> D _n-1 is established. Now, the original density data value of the original, that is, the level value D _{n-1 of} the output signal to be originally detected,
D _n and D _{n + 1} are respectively D _n-1 = 1, D _n = 10 and D _{n + 1}
= 10, the levels D _n-1 , D _n , D _{n + 1} actually detected due to the presence of the n-th defective element are D _n-1 =
It is assumed that they are given in the relationship of 1, D _n = 9, D _{n + 1} = 10. At this time, the parameter β is β = (9-1) / (1
0-1) = 8/9, and the inequality of 0 <β <1 is satisfied. Therefore, according to this method, the level D _n corrected '' is, D _n 'is given by' = D _n = 9. Therefore, the correction value of the level of the output signal of the n-th element becomes the level of the original output signal itself. On the other hand, according to the above method, the corrected level D _n 'is given by the level D _n-1 . According to the method, the corrected level D _n ′ is D _n ′ = (1 + 10) / 2 = 5.
It becomes 5. These corrected levels D _n ″ or D _n '
In FIG. 31, the solid line 54 and the chain line 5
5, 56 are shown.

The above analysis can be summarized as follows. That is, the original document density, that is, the level D _n was 10, but the presence of the defective element causes a slight output error in the output signal of the defective element, and as a result, the CCD line sensor sets the level D _{n to the} level D _n. I read it as 9. On the other hand, when the present correction method is applied, the corrected level D _n ″ is processed as data of 9. On the other hand, if the correction method or is applied, the corrected level D _n 'is processed as 1 or 5.5, and its value becomes a value farther from the true value 10. Therefore, it can be concluded that the correction method has much higher accuracy. Moreover, according to the present correction method, the level D _n ″ after correction has a value close to the level D _{n + 1} of the output signal of the next element, so that the edge portion of the pattern with black and white contrast is It is possible to eliminate unnaturalness.

As in the case of FIG. 30, this point is shown in the case where an oblique line is scanned and an image is recorded.
FIG. 32. As shown in the figure, by using this correction method, the unnaturalness of the edge portion of the correction portion is improved to the extent that it is hardly noticeable.

Similarly, as another example of reading and recording an image of a design having a contrast, the case of β> 1 is shown in FIG.
3 shows. Also in this case, according to the correction method, as shown in FIG.
The advantages described in Section 1 are similarly established. In the figure, the solid line 57 and the alternate long and short dash lines 58 and 59 respectively indicate the levels obtained by the correction method.

Next, even if the same relation of β> 1 is satisfied, the case of a flat pattern with no contrast will be examined. FIG. 34 shows this case.
Here, the original document density, that is, the levels D _n-1 , D _n , D
It is assumed that _{n + 1} are all 10, whereas the nth element is a defective element, so the CCD line sensor displays the image of the pattern as D _n-1 = 10, D _n =
It is assumed that the data is read with 9, D _{n + 1} = 10. At this time, the parameter β is β = (9-10) / (10-1)
Since 0) =-1/0 = ∞, the condition of β> 1 is satisfied. Therefore, according to the correction method, the corrected level D _n ″ is D _n ″ = D _{n + 1} = 10. In this case, the level D _n 'after correction is 1 depending on the correction method as well.
It becomes 0, which is the same as the result of the correction method. In FIG. 34, the corrected level D _n ″ is indicated by the solid line 60.
That is, in this case, the correction can be performed with high accuracy by any of the correction methods. In this case, in FIG.
33, the problem of unnaturalness at the edge portion of the recorded image, which is a problem in FIG. 33, does not matter.

Further, the case of a pattern with gentle gradation will be examined. An example of this case is shown in FIG. 35, and the parameter β is 0. Here, the original density of the original, that is, the output signal levels D _n-1 , D _n , D
_{n + 1} is 99, 100 and 101 respectively,
However, due to the presence of the nth defective element, C
The CD line sensor actually has D _n-1 = 99 and D _n = 9.
It is assumed that the data is read as 9, D _{n + 1} = 101. At this time, according to the correction method, the corrected level D _n 'is 10
The value becomes 0 and becomes the same value as the true value 100, whereas the corrected level D _n ″ becomes 99 according to the correction method. Therefore, the correction accuracy is better with the correction method. Therefore, in the correction method, there is a concern that a streak-like unevenness may occur in the pattern recorded on the film, but this is not a problem in practice. The reason is that if there are white streaky scratches in the black pattern, it will be noticeable, but in this case,
This is a case where a white streak-like flaw occurs near the boundary between both black and white images, so that the flaw is inconspicuous.

As described above, the correction method is not necessarily superior to or better than the correction method in a gentle gradation portion having no contrast, and can achieve the same level of correction accuracy, but in the case of a contrasting pattern. However, the correction accuracy is significantly higher than that of the correction method and has an excellent advantage that it hardly affects the image quality of the edge portion. Therefore, when viewed comprehensively, it can be concluded that the present correction method is a suitable method that can satisfy the requirements when applied to a device such as a plate-making scanner.

Next, the structure and operation of the image signal correction apparatus to which the correction method is applied will be described. FIG. 36 is a block diagram showing an example of such an image signal correction apparatus. The original 65 set on the glass plate 3 of the original table 2 has an image to be recorded on a film by the output device 66, and includes a picture, a line drawing, characters, tints, and the like. CCD line sensor 10 incorporated in the image reading unit 8
Is a CCD line sensor whose presence of a defective element has already been detected by using the first detection method or the second detection method described above. A / D converted image signal V ₁ output from the CCD line sensor 10
The image processing unit 63 is a part that performs shading correction on the image signal and then performs correction processing on the image signal corresponding to the defective element. Then, the image processing unit 63 stores the corrected image signal V _A in the storage device 64 such as a memory or a magnetic disk. When the image of the original document 65 is printed on a film based on these image signals V _A , the output device 66 (including the color conversion unit) may be connected to the storage device 64. The input device 20 is for inputting the position information of the defective element obtained by the first detection method or the second detection method, and the position information signal giving the position information is stored in the image processing unit 63. Stored in. The other components are the same as in the case of FIG.

FIG. 37 is a block diagram showing the structure of the image processing unit 63. The shading correction unit 68
The image signal correction processing section 69 is a section for performing shading correction of each image signal V ₁ (output signal of each element) input via the input I / F 67, and is a section for performing the above-described correction method, Two memories 71, 7
It is composed of 2 and. Other symbols 73, 74, 7
5 and 76 are a LOG converter, a noise filter, and
An output I / F and timing control 76. Of these, the memory 71 is a part that stores the image signal that has been subjected to the shading correction and also stores the image signal that has been subjected to the correction processing by the correction method. The memory 72 is a portion for storing the position information signal of the defective element described above.

Next, referring to FIG. 37 and in accordance with the flowcharts of FIGS. 38 and 39, the image signal correction processing unit 7
The operation of No. 2 will be described. First, the image signal V ₁ output from each element of the CCD line sensor 10 is stored in the memory 7
1 is stored (step S1C).

Next, the DSP 70 initializes the count value n of the counter inside the DSP 70 (this value is equal to the position number n of the element in the CCD line sensor) to 0 (step S2C), and replaces n + 1 with n. Yes (step S3C). Then, the DSP 70 determines whether or not n ≦ m (m represents the number of elements) is satisfied. If n> m, the correction process ends, and if n ≦ m, the process proceeds to the next step ( Step S4C).

The DSP 70 compares and collates the above-mentioned position information signal stored in the memory 72 in the look-up table format with the count value n, and the nth element corresponding to the count value n is a defective element. It is determined whether or not (step S5C). If it is not a defective element, the correction of the image signal is not necessary, and the process returns to step S3C. If it is determined to be a defective element, the DSP 70 proceeds to the following correction processing steps S6C to S13C.

First, the DSP 70 makes the (n-1) th,
Image signals corresponding to the nth and (n + 1) th elements are read from the memory 71, and those image signals are stored in a register (not shown) inside the DSP 70 (step S6C). Then, the DSP 70 outputs D _n-1 =
It is determined whether or not the relationship of D _n = D _{n + 1} (the symbol D indicates the level of the image signal corresponding to each element) is satisfied (step S7C), and if satisfied, the process proceeds to step S10C described later. If not satisfied, the parameter β is calculated based on the equation 5 in step S8C.

The DSP 70 examines the value of the parameter β (step S9C), sets D _n-1 to D _n ″ if B <0 (step S10C), and sets D _n to D _n if 0 ≦ β ≦ 1. ''
(Step S11C), and if β> 1, D _{n + 1 is set} to D _n ″ (step S12C). And DSP
70 determines the post-correction level D _n ″ as a new level D _n of the image signal corresponding to the n-th element, and the new level D _n is the pre-correction level D _n stored in the memory 71. Replace with _n (step S13C). As a result, the correction process of the image signal corresponding to the n-th defective element is completed. Then, step S3C again
Move on to.

(Modification of Correction Method) In the correction method, any _one of the levels D _n-1 , D _n , and D _{n + 1} for the (n-1) th, nth, and (n + 1) th elements is used. Was used depending on the value of the parameter β,
Instead of this, weighting may be applied to each of the levels D _n-1 , D _n , and D _{n + 1} , and the correction value D _n ″ may be obtained using such weighted levels. An example of such a weighting method will be described below as a correction method.

First, FIG. 40 shows the level difference value (D _{n + 1) of the} two weighting parameters W ₁ and W ₂ used in this method.
The function form for -D _n-1 ) is shown. The parameter W ₁ has a maximum value of 1 when the level difference value (D _{n + 1} −D _n−1 ) = 0, and then the level difference value (D _{n + 1} −
It decreases with the increase of D _n-1 ). On the other hand, the parameter W ₂ is the broken line 87 (each parameter W ₁ , W _{2 in} the figure).
When both are 0.5), the curve is in a line-symmetrical relationship with the parameter W ₁ with respect to the axis of symmetry.

The principle of this correction method is as follows.
Based on.

[0133]

[Equation 6]

[0134]

[Equation 7]

The weighted average level D is obtained by the processing given by the equation 6, and the weighted level D _nw is obtained by the processing given by the equation 7. Then, the weighted average value of the average level D and the level D _nw is obtained (D _n ″ = (D + D _nw ) / 2), and a new level D _n after correction is obtained using this average value D _n ″. To do. Also by this method, the output signal of the defective element can be corrected with high accuracy, and the unnaturalness at the edge portion of the recorded image can be made inconspicuous even when a document having a contrast is read and recorded. it can.

FIG. 41 is a block diagram showing the functional circuit section in the DSP 80 for executing the above correction method and the first and second memory sections 81 and 82 in the image processing section 63. However, in this figure, the memory 72 shown in FIG.
, And other functional circuit units in the DSP 80 are also omitted. The first memory unit 81 is
It corresponds to the memory 71 of FIG.

(I) As a premise, the DSP 80 has the configuration shown in FIG.
It is assumed that it is detected that the n-th element is a defective element by comparing the position information signal in the memory 72 and the count value n.

(Ii) The difference circuit section 83 receives the image signals corresponding to the (n-1) th and (n + 1) th elements from the first memory section 81 storing the shading-corrected image signal. V _Dn-1 and V _{Dn + 1} are read and a signal V _DIF giving a level difference value (D _{n + 1} -D _n-1 ) is generated. Then, the difference circuit unit 83 outputs the signal V _DIF to the second memory unit 82.

(Iii) The second memory section 82 is a memory for storing the data of both parameters W ₁ and W ₂ shown in FIG. Therefore, the second memory unit 82 receives the signal V _DIF as a signal giving the address, and outputs the signal V _DIF.
Signals V _W1 and V _W2 that give parameters W ₁ and W ₂ corresponding to the _DIF are generated, and are output to the first and second weighting units 84 and 85, respectively.

(Iv) The first weighting section 84 uses the signal V _w1 ,
Using the image signals V _Dn-1 and V _{Dn + 1} , a signal V ₀₁ that gives a weighted average level D is generated from Equation 6, and is output to the averaging unit 86. On the other hand, the second weighting unit 8
5, using the signal V _w2 and the image signal V _Dn ,
Generate a signal V ₀₂ giving a weighted level D _n ,
It is output to the average processing unit 86.

(V) The averaging processor 86 uses the two signals V ₀₁ and V _{02.
Receive, (V 01 + V 02)} / ' generates, the signal V _Dn' image signal V _Dn given by 2, to determine the new image signal corresponding to the n-th element, it Output to the first memory unit 81. As a result, the first memory unit 81 _{replaces the} uncorrected image signal V _Dn with the new image signal V _Dn '
Replace with.

In this way, the new image signal V _Dn 'after correction is _added .
Is the signal V weighted from both levels D _n-1 and D _{n + 1.}
A signal V ₀₂ obtained by weighting ₀₁ and the level D _n before correction
Given by the composite signal of.

When the method of determining the level D _n ″ developed in the description of the correction method (see equation 5) is reconsidered from the viewpoint of “weighting” developed in the above modification, equation 5 is obtained.
In the method of, each level D _n-1 , as shown in the following equation 8,
It can be said that the method is equivalent to the method of correcting the level D _n by the combination processing of D _n and D _{n + 1} .

[0144]

[Equation 8]

Here, each of the coefficients γ _{1 to} γ ₃ (combining ratio) is a weighting value, and the value of the above-mentioned parameter β and D
Depending on whether the condition of _n-1 = D _n = D _{n + 1} is satisfied, the values given in the following Table 1 are taken.

[0146]

[Table 1]

More generally speaking, the values of the respective synthesis ratios γ _{1 to} γ _{3 in} the equation 8 are expressed by the relational expressions 0 ≦ γ ₁ ≦ 1, 0 ≦ γ ₂ ≦ 1 according to the value of the parameter β. , 0 ≦ γ ₃ ≦ 1 and γ ₁ + γ ₂
+ Γ ₃ = 1 is changed within the range to satisfy each level D _n−1 ,
The composite value D _n ″ of D _n and D _{n + 1} may be determined.

The above correction method and its modification are
It goes without saying that the present invention can be applied not only to the correction of the image signal output from the CCD line sensor but also to the correction of the image signal output from the two-dimensional CCD area sensor.

Summarizing the advantages of the respective embodiments described above,
It is as follows. That is, the presence or absence of the defective element can be visually detected with high accuracy by the first detection method and. Moreover, according to the above method, the position of the defective element can also be detected visually with high accuracy. And the detection at that time can be achieved very easily because it is only necessary to visually confirm the intersection of the image of the position information and the extension line of the white flaw or the black flaw, and it is highly practical. Can be said.

Further, if the second detection method is used, the presence or absence of the defective element and its position can be specified with high accuracy by only processing the electric signal without outputting the output film by the output machine. can do. The fact that all can be achieved by electrical processing has the advantage that the detection time can be significantly shortened compared to the first detection method.
In addition, the work load on the operator can be significantly reduced.

By using the first and second detection methods described above, it is possible to accurately detect a defective element at a level that could not be detected in the inspection stage at the time of manufacturing and shipping an image sensor represented by a CCD sensor. Can be detected with. Therefore, the first or second detection method,
The present invention opens the way to the application of plate-making equipment such as a scanner to which this method is applied as a device for detecting a defective element at the stage of manufacturing and shipping an image sensor.

If the above correction method is used,
Even if a CCD sensor including a defective element is used, it is possible to prevent the occurrence of black scratches and white scratches without deteriorating the image quality of recorded images. This eliminates the need to select the CCD sensor on the image recording device side, and
This has the effect of significantly improving the yield of acceptance of the CD sensor.

[0153]

The invention according to claim 1 has an effect that the presence of a defective element can be visually and accurately detected.

The invention according to claim 2 has an effect that the detection of the defective element and the specification of the array position of the defective element can be realized visually with high accuracy.

The invention according to claims 3 and 4 has an effect that the presence of a defective element can be detected with high accuracy by electrical processing of an output signal.

The invention according to claim 5 has an effect that position information of a defective element can be determined with high accuracy by electrical processing of an output signal.

According to a sixth aspect of the present invention, an undesired image having stripes such as white scratches and black scratches is not generated in an image recorded by an image signal output from the image sensor,
Moreover, there is an effect that the image signal corresponding to the defective element can be corrected with high accuracy so as to enable the recording of a high quality image without causing unnaturalness in the edge portion of the image to be recorded. Therefore, the present invention can make it possible to apply such an image sensor to an image recording apparatus even if the image sensor has a defective element.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing an example of an output voltage of a normal element.

FIG. 2 is an explanatory diagram showing shading correction of an output voltage of a normal element.

FIG. 3 is an explanatory diagram showing an example of output characteristics of an element that is considered to cause white scratches in the case of a positive film.

FIG. 4 is an explanatory diagram showing an example of a case where the output characteristic of a defective element that causes a white defect is shading-corrected.

FIG. 5 is an explanatory diagram showing an example of output characteristics of elements that are considered to cause black scratches in the case of a positive film.

FIG. 6 is an explanatory diagram showing an example of a case where the output characteristic of a defective element that causes a black defect is shading-corrected.

FIG. 7 is an explanatory diagram showing an example of a result of reading an original image with a CCD sensor having a defective element and printing it on a positive film.

FIG. 8 is a block diagram schematically showing a configuration of an inspection device for defective elements of a CCD line sensor.

FIG. 9 is a flowchart showing a procedure of a detection method.

FIG. 10 is a flowchart showing a procedure of a detection method.

FIG. 11 is an explanatory diagram showing an overall view of a result of printing an image that gives positional information of each element on a film together with an image of a gradation original.

FIG. 12 is an explanatory diagram showing a partially enlarged view of a result of printing an image giving positional information of each element on a film together with an image of a gradation original.

FIG. 13 is an explanatory diagram showing an overall view of a result of printing an image of a gradation original document and an image giving position information of each element by enlarging it twice in the main scanning direction.

FIG. 14 is an explanatory diagram showing a partially enlarged view of a result of printing an image of a gradation original document and an image giving position information of each element by enlarging it twice in the main scanning direction.

FIG. 15 is a flowchart showing a procedure of a detection method.

FIG. 16 is a flowchart showing a procedure of a detection method.

FIG. 17 is a flowchart showing a procedure of a detection method.

FIG. 18 is a block diagram showing an example of a detection device for detecting a defective element by image data processing.

FIG. 19 is an explanatory diagram showing the criteria for detecting defective elements.

FIG. 20 is an explanatory diagram showing a detection state of elements when the original degree of a gradation original is changed stepwise.

FIG. 21 is a flowchart showing the procedure of the second detection method.

FIG. 22 is a flowchart showing the procedure of the second detection method.

FIG. 23 is a flowchart showing the procedure of the second detection method.

FIG. 24 is an explanatory diagram showing a positional relationship of each element of the CCD line sensor.

FIG. 25 is a flowchart showing a procedure of a calculation method.

FIG. 26 is a flowchart showing a procedure of a calculation method.

FIG. 27 is a flowchart showing a procedure of a modified example of the second detection method.

FIG. 28 is an explanatory diagram schematically showing an application example of the correction method.

FIG. 29 is an explanatory diagram schematically showing an application example of the correction method.

FIG. 30 is an explanatory diagram which points out a problem that an edge portion of an image recorded on a film is conspicuous.

FIG. 31 is an explanatory diagram showing an example of applying the correction methods to in the case of a design having a contrast.

FIG. 32 is an explanatory diagram showing an effect of the correction method when an oblique line is scanned and an image is recorded.

FIG. 33 is an explanatory diagram showing another example of applying the correction methods 1 to 3 in the case of a contrasting pattern.

FIG. 34 is an explanatory diagram showing an example of applying the correction methods to in the case of a flat pattern with no contrast.

FIG. 35 is an explanatory diagram showing an example of applying the correction methods to in the case of a pattern with a gentle gradation.

FIG. 36 is a block diagram showing an example of an image signal correction apparatus to which a correction method is applied.

FIG. 37 is a block diagram showing a configuration of an image processing unit in the image signal correction apparatus.

FIG. 38 is a flowchart showing the operation of the image signal correction processing section.

FIG. 39 is a flowchart showing the operation of the image signal correction processing section.

FIG. 40 is an explanatory diagram showing a function form of a weighting parameter with respect to a level difference value.

FIG. 41 is a block diagram showing each functional circuit unit of the DSP that executes the correction method and the first and second memory units.

[Explanation of symbols]

1 Planar Input Scanner 2 Platen 4 Gradation Document 5 Halogen Lamp 10 CCD Line Sensor 13 Control Unit 14 Transmitted Light 20 Input Device 21 Image Processing Unit 23 Memory 31 Output Scanner 32 Film 33 First Printing Area 34 Second Printing Area 42 Image Processing unit 47 DSP 63 Image processing unit 65 Original document V ₁ output signal V _PI position information signal V _A Corrected image signal

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G01R 31/26 FG06T 1/00 G06K 9/20 320 C H01L 21/66 J 7630-4MH H04N 1 / 028 Z 1/40 H04N 1/40 G

Claims (6)

[Claims] 1. An image sensor having a plurality of elements that perform photoelectric conversion is made to enter light while changing the amount of incident light, the output signals of the plurality of elements are measured, and an image given by each output signal is used as a sensitive material. A method of detecting a defective element of an image sensor, comprising: recording and detecting whether or not the image sensor has a defective element that causes an abnormality in output characteristics with respect to the incident light amount based on the recorded image. 2. At the time of the recording, together with the images of the output signals of the plurality of elements, an image of position information indicating the arrangement position of each element in the image sensor is also recorded on the photosensitive material, The image according to claim 1, wherein by comparing both recorded images of the information and the output signal, the array position of the defective element is specified in addition to the detection of the presence or absence of the defective element. Method of detecting defective element of sensor. 3. A step of: (a) injecting a certain amount of light into an image sensor having a plurality of elements for photoelectric conversion, and (b) a defective element exhibiting abnormality in a region having an output characteristic with respect to the amount of incident light. A step of detecting whether or not the element exists in the plurality of elements based on the respective output signals of the plurality of elements; and (c) changing the light quantity of the incident light in the step (a), And a step of repeating both the steps (a) and (b) using light having a light amount. 4. The step (b) includes: (b-1) determining whether or not the level of each output signal is within an allowable range determined according to the incident light amount; and (b-2) ) Determining that the corresponding element is the defective element when the level is not within the permissible range.
A method for detecting a defective element of the image sensor described. 5. In the step (b), when the element to be detected is the defective element, the array position of the defective element in the image sensor is specified, and then the positional information for giving the array position is provided. 5. The method for detecting a defective element of an image sensor according to claim 3, further comprising a step of recording 6. Among defective image signals which give an image of a document read by an image sensor, among a plurality of elements for performing photoelectric conversion included in the image sensor, a defective element having an abnormal output characteristic with respect to an incident light amount is selected. An image signal correction method for correcting a corresponding image signal, wherein an image signal corresponding to the defective element and an image signal corresponding to the element located in the vicinity of the defective element are combined, and obtained by the combining. An image signal correction method, characterized in that the image signal is determined as a new image signal corresponding to the defective element.