JP2002281391A - Imaging system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize an imaging system with high performance not to virtually generate degradation of image quality due to increase of defects of pixels over aging without adding a special device such as a light shielding means and a white light imparting means to structure. SOLUTION: A system controller 112 is provided with a defect data detecting part 112a to detect pixel defect data based on detection results of every main photography and a defect compensation control part 112b to perform pixel defect compensation processing to a signal to be obtained from a CCD 105 in the case of the main photographing. In defect detection by the defect data detecting part 112a, candidates for defects are detected by analyzing an image, for example, in the case of every main photographing and only a candidate for the defects to satisfy prescribed conditions in a plurality of times of detection is further determined as defect pixel data. And the defect pixel data are additionally registered in an EEPROM 118.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image pickup apparatus, and more particularly to an image pickup apparatus having a pixel defect compensation function.

[0002]

2. Description of the Related Art Imaging devices such as video cameras have been widely used. In recent years, electronic still cameras that mainly capture and record still images have also come into widespread use especially as digital cameras, and video movies mainly for recording moving images also have a still image capturing and recording function. Long-time exposure, which is mainly used for shooting still images, increases the exposure time by increasing the charge accumulation time in the image sensor, thereby enabling shooting without using auxiliary lighting such as a strobe even under low illumination. It is known as a technology.

On the other hand, an image pickup device such as a CCD has a dark output due to the presence of a so-called dark current, which is superimposed on an image signal, thereby deteriorating the image quality. If there is a pixel with a large dark output level, it is called a pixel defect,
It is widely practiced to supplement information using output information of neighboring pixels without using output information of the pixel. In the present specification, such a complementing process is referred to as pixel defect compensation. A predetermined (often 1/60 in NTSC) which is often determined on the premise of driving a moving image at the used frame rate
The dark output is evaluated at a standard exposure time of seconds or a predetermined margin based on the predetermined margin (for example, 1/15 second when the margin is 4 times), and a pixel having a large level is regarded as a defective pixel. Apply the pixel defect compensation.

Further, Japanese Patent Application Laid-Open No. 06-038113 discloses a technique for improving that a pixel defect involves temperature dependence and aging, so that it is not sufficient to evaluate a defective pixel just before shipment from a factory. It is known. That is, in this publication, the light receiving surface is shielded by closing the iris immediately after the power is turned on, and the C
A technique for detecting a defective pixel by evaluating a CD dark output and performing defect compensation is described. According to this method, a so-called white defect caused by dark charge can be detected by the camera alone, but a black defect cannot be detected.

On the other hand, for example, Japanese Patent Application Laid-Open No. 06-006
No. 685 discloses white light (reference light) on an imaging surface.
If the providing means is used, the detection of the black defect is similarly performed,
It describes that this can be performed by a camera alone.

[0006]

By the way, in the camera device (as a completed product after shipment from the factory) as in the technique disclosed in the above-mentioned publication, it is necessary to detect a defective pixel every time the power is turned on to perform defect compensation. This can be a drawback in operating the camera. That is, in order to detect a defective pixel, an iris, a mechanical shutter, or the like is once operated to block light, and then an imaging operation (charge accumulation, transfer, signal reading) is performed, and a defective pixel address obtained by data analysis is obtained. Since it is necessary to execute a series of sequences of writing data to the memory, a certain period of time (for example, several hundred ms to several s) is required. During this time, the user of the camera needs to wait for a considerable amount of time, so he cannot understand why it takes a lot of time to enter the photographable (standby) state, which may cause confusion such as determining that the camera is out of order. There is. Even if the display function is devised to avoid misunderstanding, there is a high possibility that a photographing opportunity will be missed before photographing becomes possible. Needless to say, this problem is particularly serious in a still image camera which generally has a high demand for quick shooting performance.

[0007] The biggest problem is that the technique disclosed in the above-mentioned publication requires a light shielding means and a white light applying means as an essential device. Although the light-blocking means can be commonly used as an essential exposure control device such as an aperture and a shutter, it is still difficult to adopt a configuration in which the exposure control device is omitted particularly for a simple configuration. In order to detect a black defect, it is necessary to add a white light providing device, which has not been required in the related art, to the configuration, and there has been a very serious problem that the size and cost of the camera device are increased.

By the way, the temperature characteristics of pixel defects can be determined by using other known methods such as incorporating a temperature sensor in the camera and storing defect data for each temperature when the camera is shipped. It is also possible to avoid a method requiring a long time lag. On the other hand, it goes without saying that in order to apply the defect compensation to the deterioration over time of the pixel defect of the image sensor after the product is shipped, the defect detection by the camera itself is inevitable.

However, it is not generally considered that such a time-dependent deterioration of a pixel defect increases rapidly in a short time. For example, one pixel is generated one day due to the influence of cosmic rays or natural radioactivity, and several weeks later. It occurs with a very low probability, such as the occurrence of one pixel. Of course, once it occurs, it is virtually impossible to disappear, so even if it is left at one pixel level, a defect will be permanently manifested if left unchecked. However, if the new defect detection function in the camera can detect this shortly and apply compensation processing, the probability of causing damage to the user that the defect becomes apparent becomes extremely small, and However, even if a defect becomes apparent, it is limited to only one time, and the disadvantage is practically negligible.

The present invention has been made in consideration of the above circumstances, and has as its object to solve the above-described problems caused by detecting defect data for pixel defect compensation.
It is an object of the present invention to provide a high-performance image pickup apparatus which does not cause any image quality deterioration due to an increase in pixel defects over time without adding a special apparatus to the configuration and without causing a photographing time lag.

[0011]

In order to solve the above-mentioned problems, an image pickup apparatus according to the present invention comprises an image pickup device, an image pickup optical system for inputting a subject image to the image pickup device, and a defective pixel address of the image pickup device. Storage means for registering as defect data;
A defect compensating unit that performs a compensation process based on defect data registered in the storage unit with respect to an output of the image sensor, based on defect data, and a normal exposure state different from either a light-shielded state or a uniform light input state. Defect data detecting means for detecting the defective pixel address by analyzing the output of the imaging element; and defect data registered in the storage means based on the defect data newly detected by the defect data detecting means. And a defect data management means for updating.

In this image pickup apparatus, the output of the image pickup element in a normal exposure state different from either a light-shielded state or a uniform light input state is analyzed, for example, at the time of each main image pickup, to detect a defective pixel address. It is configured to perform. By providing the function of analyzing the output of the image sensor in the normal exposure state for detecting a defective pixel address in this way, a late defect can be detected without providing a special device such as a light shielding unit or a white light providing unit. It is possible to do.

In this case, the defect data detecting means includes:
A defect candidate data detecting means for detecting a defective pixel candidate address by analyzing an output of the imaging element in the normal exposure state; and a plurality of times of the result of the defect candidate data detection executed by the defect candidate data detecting means for the plurality of times. It is preferable to use a configuration having defect determining means for determining a defective pixel address by performing evaluation based on a comprehensive standard covering:

[0014] Thereby, for example, the image is analyzed each time a main image is picked up to detect a defect candidate, and only a defect candidate that satisfies a predetermined condition in a plurality of detections is registered as a defect by address registration. Defect detection can be performed with high accuracy without requiring any special means for detection.

Regarding the black defect candidate pixel, 1) regarding the output of the image sensor in the normal exposure state, when the level of the pixel is lower than a neighboring pixel by a predetermined level or more, this is detected as a black defect candidate pixel. Standards,
In order to further improve the accuracy, 2) regarding the output of the image sensor in the normal exposure state, when the level of the pixel is smaller than a neighboring pixel by a predetermined level or more and is almost 0, this is detected as a black defect candidate pixel. It is preferable to use a reference.

Similarly, for the white defective pixel, 1)
Regarding the output of the image sensor in the normal exposure state, when the level of the pixel is higher than a neighboring pixel by a predetermined level or more, the standard for detecting this as a white defect candidate pixel, and in order to further improve the accuracy, 2) normal exposure Regarding the output of the image sensor in the state, it is preferable to use a criterion for detecting the pixel as a white defect candidate pixel when the level of the pixel is higher than a neighboring pixel by a predetermined level or more and is substantially at a saturation level.

[0017] Preferably, the defect determining means preferably determines, as a defective pixel, a defect candidate detected as the same type of defect candidate in all of the plurality of times of defect candidate data detection executed by the defect candidate data detecting means. Or, more preferably, those that are detected as the same type of defect candidate for all of the plurality of times of defect candidate data detection, and that the level of a pixel adjacent to the defect candidate pixel has a different value in the plurality of times. A configuration determined as a defective pixel can be used.

This makes it possible to improve the accuracy of defect detection. In the latter case, the defective pixel determination condition is that the pattern of the subject image has changed in a plurality of detections of defect candidate data. Therefore, it is possible to solve the problem that a specific pattern is erroneously detected as a defect.

The problem of erroneously detecting a picture of a specific subject image as a defect as described above is that the zoom operation of the photographing optical system is performed from the time when the previous defective pixel candidate address is detected until the time when the next defective pixel candidate address is detected. Can be effectively avoided by a configuration in which the detection of the defective pixel candidate address by the defective candidate data detection means is stopped when the processing is not performed.

In the case where neighboring pixel data is used for defect detection, in order to prevent erroneous detection, if there is already a registered defective pixel among neighboring pixels, the neighboring pixel is deleted. It is desirable to detect a defective pixel address by regarding a pixel other than the registered defective pixel as a neighboring pixel.

[0021]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a digital camera according to an embodiment of the present invention.

In the figure, reference numeral 101 denotes an image pickup lens system as a variable power optical system including a zoom lens, 102 denotes a lens driving mechanism for driving the lens system 101, and 103 denotes a lens system 101.
An exposure control mechanism 104 for controlling the aperture and shutter devices of the camera; 104, a low-pass and infrared cut filter;
05 is a CCD color image sensor for photoelectrically converting a subject image, 106 is a CCD for driving the image sensor 105
A driver 107, a pre-processing circuit including an A / D converter, etc., a digital processing circuit 108 for performing various digital arithmetic processing such as γ conversion,
9 is a card interface, 110 is a memory card,
Reference numeral 111 denotes an LCD image display system. In the figure, reference numeral 112 denotes a system controller (CPU) for integrally controlling each unit; 113, an operation switch system including various switches; 114, an operation display system for displaying an operation state, a mode state, and the like; Is a lens driver for controlling the lens driving mechanism 102, 116 is a strobe as a light emitting means, 117 is an exposure control driver for controlling the exposure control mechanism 103 and the strobe 116, and 118 is a memory for storing various setting information and the like. Non-volatile memory (EEP
ROM).

In the camera according to the present embodiment, the system controller 112 performs overall control, and controls the exposure control mechanism 103 and the CCD driver 106 to drive the CCD image pickup device 105 for exposure (charge accumulation). And read out the signal, A / D convert the signal through the pre-processing circuit 107, take it into the digital process circuit 108, perform various signal processing in the digital process circuit 108, and then perform the card interface 1
09 to the memory card 110.

The system controller 112 includes:
As shown in the figure, a defect data detection unit 112a for detecting pixel defect data based on a detection result at each main imaging, and a pixel defect compensation process for a signal obtained from the CCD 105 at the time of main imaging. And a defect correction control unit 112b.

The pixel defect compensation processing is performed by the defect correction control unit 11
Based on the instruction from 2b, the process is executed in the digital process circuit 108 based on the address data of the defective pixel stored in the EEPROM 118 regarding the existing (latest) defect (hereinafter referred to as a registered defect). In the initial stage (when the camera is shipped from the factory), defect data acquired in the manufacturing adjustment process is registered as a registered defect.

The defect detection by the defect data detecting unit 112a is performed, for example, by analyzing an image at each actual imaging to detect a defect candidate, and furthermore, detecting a defect candidate satisfying a predetermined condition in a plurality of detections. This is performed by registering an address as pixel data.

Hereinafter, camera control by the system controller 112 will be described focusing on processing directly related to detection and compensation of pixel defects according to the present invention. However, in this camera, the digital processing of the signal level obtained from the CCD 105 is performed with 8 bits (0 to 255). Except for the parts specially described later, the description will be made assuming normal temperature.

Prior to photographing, an exposure time necessary for photographing is set based on a manual setting or a photometric result. Next, it waits for a shooting trigger command of the main imaging, and upon receiving the command, performs exposure based on a predetermined exposure control value, reads an imaging signal, temporarily stores the data in a buffer memory,
On the one hand, the defective pixel candidate is held for detecting a defective pixel candidate to be described later, and on the other hand, predetermined signal processing is immediately performed and then recorded on the memory card 110. At this time, the registered defective pixel is accompanied by pixel defect compensation. After the defect compensation, the video signal processing up to the recording is appropriately used according to its necessity. It is known per se, for example, color balance processing, conversion into a luminance-color difference signal by matrix operation or its inverse conversion processing, Examples include false color removal or reduction processing by band limitation, various nonlinear processing represented by γ conversion, various information compression processing, and the like.

In the camera of the present embodiment, the well-known "complementation of a pixel in which a defect address is registered with a neighboring pixel" is employed as a defect compensation. Among them, the four pixels closest to the defective pixel: In the case of an RGB Bayer array, for example, four G pixels adjacent to each other diagonally in four directions for G, and four pixels directly adjacent to R (or B) in four directions, up, down, left and right Four R (or B) located next to each other with one G in between
Pixel), the average value of four pixel information is substituted.

The present camera detects a defect candidate using the image signal data stored in the buffer memory at an appropriate time after the recording is completed. A defective pixel is determined based on a reference, and the pixel address of the registered defect in the EEPROM 118 is additionally updated based on the defective pixel. As described above, the present camera is configured to detect the defective pixel address by analyzing the output of the CCD 105 at the time of the main imaging different from either the light shielding state or the uniform light input state.

The operation of detecting a defect candidate will now be described with reference to the flowchart of FIG.

First, the output level of each pixel is analyzed using the image pickup signal stored in the buffer memory (step S101), and the level becomes 5 (saturation level 2).
Pixels equal to or less than 55% (approximately 2% of 55) are selected as tentative candidates for a black defect, assuming that the output is nearly zero (step S1)
02, S103). The level is 250 (about 98%)
For the above pixels, the output is almost saturated level (255)
And is selected as a temporary candidate for a white defect (steps S106 and S107).

Next, for each of the provisional black defect candidates,
The level of the pixel and the neighboring four pixels related to the pixel are compared, and a black defect temporary candidate pixel smaller than any of the neighboring four pixels and having a difference of 50 (about 20%) or more is determined as a black defect candidate. (Step S104,
S105). That is, by examining the relative level difference between neighboring pixels that should have the same output level, a black defect candidate pixel is determined from among the black defect temporary candidate pixels whose output is almost zero. Is done.

Similarly, for each of the temporary white defect candidates, the level of the pixel is compared with that of the four neighboring pixels related to the pixel, and is greater than any of the neighboring four pixels, and the difference is 50 (about 20). %) Or more are detected as white defect candidate pixels (step S108,
S109). In other words, by examining the relative level difference between each of the neighboring pixels that should have the same output level, the black defect candidate pixel is output from the white defect temporary candidate pixels whose output is almost at the saturation level. It is determined.

However, in the above-described defect candidate detection operation, pixels that have already been registered as defective pixels are excluded from candidates for candidate detection processing (they are included in the target, and duplication is taken into account during registration at a later stage). Is also good). If a pixel already registered as a defective pixel is included in the neighboring pixels, the processing in S104 and S108 is omitted in order to correctly determine the relative level difference from the defective temporary candidate pixel. Do it.

Processing for selecting temporary candidate pixels for black defects and white defects according to the level of each pixel (S102, S10
3, S106, S107) may be omitted, and defect candidates may be determined based on a relative level difference between all the temporary candidate pixels and neighboring pixels corresponding to the temporary candidate pixels.

In this way, by acquiring the pixel address of the defect candidate each time the main imaging is performed, the defect candidate detection processing is executed a plurality of times under the normal exposure condition.

Next, referring to the flowchart of FIG.
The defect determination processing will be described. This defect determination process
The five most recent defect candidate detections (five shootings in this example)
The defective pixel is determined based on the result of (1), and the address of the detected defective pixel is additionally registered in the EEPROM 118. Specifically, a pixel detected as a black defect candidate in all of the last five defect candidate detections is determined as a black defect, and a pixel detected as a white defect candidate in all of the last five defect candidate detections is determined as a white defect. Such a detected defective pixel is referred to as a detected defective pixel.

That is, after the completion of the defect candidate detection processing, it is first determined whether or not the five most recent defect candidate detections have been completed (step S111). When the five latest defect candidate detections have been completed and it is confirmed that the results of the five defect candidate detection processes have been accumulated (Y in step S111).
ES) The pixel address determined as a black defect candidate in all the defect candidate detections is obtained by referring to the defect candidate detection results for each of the five times, and the pixel address excluding the overlap with the registered black defect pixel is obtained. A new black defect detection pixel address is additionally registered in the EEPROM 118 (steps S112 and S113).

Similarly, with respect to the white defect candidate pixel, the pixel address determined as the white defect candidate in all the defect candidate detections is obtained by referring to the defect candidate detection results for each of the five times, and the registered white defect candidate is obtained. The pixel address excluding the overlap with the pixel is additionally registered in the EEPROM 118 as a new white defect detection pixel address (step S11).
4, S115).

If the camera is fixed on three legs and the same scene is continuously photographed, there is a possibility that a white point or a black point of the subject is erroneously detected as a pixel defect. To cope with this, the following improvement examples can be applied.

(Improvement Example 1) This is to execute the defect candidate detection at an interval of, for example, (shortest) 2 hours by a timer built in the camera. FIG. 4 shows a processing procedure in this case.

FIG. 4 shows an imaging / recording operation at the time of actual imaging. First, prior to shooting, the exposure time required for shooting is set based on the manual setting or the photometric result, and variable focus control corresponding to the zoom operation performed by the photographer as needed by manual or electric drive system After the focus control process is performed, when a shooting trigger command for main imaging is received, exposure based on a predetermined exposure control value is performed, and an imaging signal is read from the CCD 105 (step S121). The image signal is input to the buffer of the digital process 108 after A / D conversion, where the image signal is subjected to defect compensation processing based on registered defect data, and then is recorded on the memory card 110 through various other image processing ( Step S122). In parallel with this, it is determined whether or not the actual imaging is performed two hours after the previous detection of the defect candidate (step S123). If so, the defect candidate detection process described with reference to FIG. 2 is executed. (Step S124) If two hours have not yet elapsed, the execution of the defect candidate detection processing is stopped, and is postponed until the main imaging is performed two hours after the previous defect candidate detection. By providing such a time span, when the same scene is continuously photographed, the defect candidate detection performed at that time can be performed only once. The possibility of erroneously detecting a defect can be greatly reduced.

(Improvement Example 2) The memory card 110
For example, the above-described candidate detection may be performed when the recording medium is exchanged. In other words, the configuration is such that candidate detection is performed only at the time of the first main imaging after the recording medium is replaced. When the recording medium is replaced, the shooting location is changed from the previous shooting, or at least the shooting composition is often changed even in the same shooting location. Therefore, the same effect as in the first improvement example is expected. I can do it.

In each of the improved examples 1 and 2, the possibility of erroneous detection is reduced, but the possibility exists significantly. Then, the following improvement examples are considered.

(Improvement Example 3) In the zoom camera, the above-described candidate detection is executed on condition that the zoom operation is performed, and the zoom operation is performed from the previous candidate detection time to the next candidate detection time. If there is no candidate, the current candidate detection is stopped. FIG. 5 shows an operation example in this case.

First, prior to photographing, an exposure time required for photographing is set based on manual setting or a photometric result, and a variable corresponding to a zoom operation by a photographer performed as needed by a manual or electric drive system. After the focus control / focusing control process is performed, when a shooting trigger command for main imaging is received, exposure based on a predetermined exposure control value is performed, and reading of an imaging signal from the CCD 105 is performed (step S131). The image signal is input to the buffer of the digital process 108 after A / D conversion, where the image signal is subjected to defect compensation processing based on registered defect data, and then is recorded on the memory card 110 through various other image processing ( Step S132). In parallel with this, it is determined whether or not a zoom operation has been performed in step S131 at the time of the next shooting (step S13).
3) If so, the defect candidate detection process described with reference to FIG. 2 is executed (step S134), but if no zoom operation is performed, the execution of the defect candidate detection process is stopped and the zoom operation is performed. This is postponed until the next main imaging is performed. (In other words, the defect candidate detection processing is executed in response to the zoom operation.) When the zoom operation is performed, the optical image formed on the CCD 105 by the previous photographing is replaced with the optical image formed on the CCD 105 by the previous photographing. Since a picture different from an image is almost always formed, the possibility of erroneous detection can be reduced as compared with the method of providing a time span as in the first and second improved examples. In addition, by appropriately increasing the number of defect candidate detection processes to be performed until the defect determination process is performed, the possibility of erroneous detection can be substantially reduced to zero.

In this example, as described above, when the zoom operation is not performed, the execution of the defect candidate detecting process is stopped. This is intended to shorten processing time and save power by preventing unnecessary processing from being executed. However, it is not always necessary to stop the processing, and a plurality of defective pixels can be determined through a zoom operation when determining a defect. What is necessary is just to use the detection result of the candidate address.

(Improvement Example 4) In each candidate detection, data of neighboring pixels is also recorded, and a black (white) defect candidate is detected in a plurality of (for example, at least two times of the previous and current) candidates. When any of the data of the neighboring pixels is significantly different between the previous and current
5 = about 10% or more), the defect candidate is determined as a black (white) defect.

FIG. 6 shows an example of the defect candidate detection processing in this case, and FIG. 7 shows an example of the defect determination processing.

As shown in FIG. 6, in the defect candidate detection processing, first, the output level of each pixel is analyzed using the image pickup signal stored in the buffer memory (step S141), and the level becomes 5 (saturation). About 2 of level 255
%) Or less is selected as a provisional candidate for a black defect, assuming that the output is almost close to 0 (steps S142 and S1).
43). Pixels having a level of 250 (about 98%) or more are selected as temporary candidates for white defects, assuming that the output is almost close to the saturation level (255) (step S1).
46, S147).

Next, for each of the provisional black defect candidates,
The level of the pixel and the neighboring four pixels related to the pixel are compared, and a black defect temporary candidate pixel smaller than any of the neighboring four pixels and having a difference of 50 (about 20%) or more is determined as a black defect candidate. And the pixel address of the black defect candidate is recorded together with the level and address of each of the four neighboring pixels corresponding thereto (steps S144 and S144).
145).

Similarly, for each of the temporary white defect candidates, the level of the pixel is compared with that of the neighboring four pixels related to the pixel, and is larger than any of the neighboring four pixels, and the difference is 50 (approximately 50). (20%) or more are detected as white defect candidates, and the pixel address of the white defect candidate is recorded together with the level and address of each of the four neighboring pixels corresponding thereto (step S148,
S149).

However, as in the case of FIG. 2, pixels that have already been registered as defective pixels are excluded from candidates for candidate detection processing (they are included in the target, and duplication is taken into account during registration at a later stage). Is also good). If a pixel already registered as a defective pixel is included in the neighboring pixels, the processing in S144 and S148 is omitted in order to correctly determine the relative level difference from the defective temporary candidate pixel. Do it.

Further, a black defect is determined according to the level of each pixel.
Process of selecting a temporary candidate pixel of white defect (S142, S14)
3, S146, and S147) are omitted, and all pixels are set as temporary candidates, and a defect candidate is determined for each of the temporary candidate pixels based on a relative level difference between the pixel and a neighboring pixel corresponding thereto. They may be recorded together.

As shown in FIG. 7, after the completion of the defect candidate detection processing, first, it is determined whether or not a predetermined number of (for example, five, at least two) defect candidate detections have been completed. Is determined (step S151). When the defect candidate detection for the plurality of times is completed and it is confirmed that the results of the defect candidate detection processing for the plurality of times are accumulated (YES in step S151), the defect candidate detection result for each of the plurality of times is referred to. Thus, the pixel determined as the black defect candidate in all the defect candidate detections is obtained (step S15).
2) The following processing is performed for each pixel.

That is, first, for each of the pixels, a level comparison between neighboring pixel data recorded in each of the plurality of candidate detections is performed (step S15).
3) It is determined whether at least one pixel having a level difference exists in the data of the four neighboring pixels (step S154). If there is a pixel having a level difference in the neighboring pixel data, it is determined that the image of the subject that has been the imaging target is different between the plurality of candidate detections,
The pixel of the defect candidate is E as a black defect detection pixel address.
It is additionally registered in the EPROM 118 (step S15)
5).

Similarly, with respect to the white defect candidate pixels, the pixels determined as white defect candidates in all the defect candidate detections are obtained by referring to the defect candidate detection results for a plurality of times (step S156). The following processing is performed on the pixel.

That is, first, for each of the pixels, a level comparison between neighboring pixel data recorded in each of the plurality of candidate detections is performed (step S15).
7) It is determined whether or not even one pixel having a level difference exists in the data of the four neighboring pixels (step S158). If there is a pixel having a level difference in the neighboring pixel data, it is determined that the image of the subject that has been the imaging target is different between the plurality of candidate detections,
The pixel of the defect candidate is E as a white defect detection pixel address.
It is additionally registered in the EPROM 118 (step S15)
9).

Also in the improvement example 4, the possibility of erroneous detection can be reduced to practically zero by appropriately increasing the number of defect candidate detection processes to be performed before performing the defect determination process.

Further, with respect to the improved examples 3 and 4, the image pickup output signals obtained continuously in the moving image mode such as when used as an EVF (every frame or frame) are not used every time the still image is shot as described above. It is possible to use the frame (timely sampling the frame), and in such a case, it is possible to detect a pixel defect (candidate detection and determination by a test) within a very short time in a normal use situation. .

The EEPR of the address of the detected defective pixel
At the time of additional registration in the OM 118, if the already registered defective pixel is not excluded from the candidate detection processing,
What is necessary is just to register what removed the registration defect from the registration defect. In such a case, the data of the detected defect may be simply replaced with the existing registration data. However, when such a configuration is adopted, if a detection error occurs due to some cause such as noise at the time of new detection, a pixel which has been defective due to deterioration over time at the time of shipment from the factory or a non-defective pixel is determined. And the defect may be exposed. On the other hand, in the above-described embodiment, since the defect obtained by removing the overlap with the existing registered defect from the detected defect is added to the existing data, it is possible to prevent the once registered defect from re-emerging. ing.

In addition, if the supplementary explanation is given regarding the quantization level of the AD converter used in the above embodiment, actually,
Considering the existence of the error characteristic of the AD converter hardware and the fact that even if it does not exist, in principle, the quantization error near the minimum quantization level is relatively equivalent to 100%, The AD converter used for the actual quantization with respect to the form is, for example, 10 which is larger than the number of quantization bits (8 bits in the embodiment) of the image processing system.
It is more preferable to use a bit or about 12 bits (or more), so that the effects of errors can be sufficiently reduced in the calculation of each of the above arithmetic expressions.

It is to be noted that the respective improved examples described in the above embodiments can be used in appropriate combinations. Further, even if a plurality of candidate detections are continuously and automatically executed while driving the zoom, it is possible to prevent erroneous detection from occurring if the normal exposure state is different from either the light-shielded state or the uniform light input state. Becomes possible.

The embodiments described above include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some components are deleted from all the components shown in the embodiment, the problem described in the column of the problem to be solved by the invention can be solved, and the effects described in the column of the effect of the invention can be solved. Is obtained, a configuration from which this configuration requirement is deleted can be extracted as an invention.

[0066]

As described above, according to the present invention,
It is possible to realize a high-performance imaging apparatus that does not substantially cause image quality deterioration due to an increase in pixel defects over time without adding a special apparatus to the configuration.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an electronic camera according to an embodiment of the present invention.

FIG. 2 is an exemplary flowchart illustrating the procedure of a defect candidate detection process performed by the electronic camera of the embodiment.

FIG. 3 is an exemplary flowchart illustrating the procedure of a defect determination process performed by the electronic camera of the embodiment.

FIG. 4 is an exemplary flowchart for explaining the operation of the electronic camera according to the embodiment at the time of actual imaging;

FIG. 5 is an exemplary flowchart for explaining another example of the operation of the electronic camera according to the embodiment at the time of actual imaging;

FIG. 6 is an exemplary flowchart illustrating another example of the defect candidate detection process performed by the electronic camera of the embodiment.

FIG. 7 is an exemplary flowchart illustrating another example of the defect determination processing performed by the electronic camera of the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 ... Lens system 102 ... Lens drive mechanism 103 ... Exposure control mechanism 104 ... Filter system 105 ... CCD color image sensor 106 ... CCD driver 107 ... Preprocess circuit 108 ... Digital process circuit 109 ... Card interface 110 ... Memory card 111 ... LCD image Display system 112: System controller (CPU) 112a: Defect data detection unit 112b: Defect compensation control unit 118: EEPROM

Claims (11)

[Claims] 1. An image pickup device, an image pickup optical system for inputting a subject image to the image pickup device, a storage unit for registering a defective pixel address of the image pickup device as defect data, and a storage for an output of the image pickup device Defect compensation means for performing compensation processing based on neighboring pixel data based on the defect data registered in the means, and analyzing the output of the image sensor in a normal exposure state different from either a light-shielded state or a uniform light input state. Defect data detection means for detecting a defective pixel address; and defect data management means for updating the defect data registered in the storage means based on the defect data newly detected by the defect data detection means. An imaging device characterized by the above-mentioned. 2. The defect data detecting means for detecting a defective pixel candidate address by analyzing an output of an image pickup device in the normal exposure state, and the defect data detecting means executes the defect candidate data detecting means. And a defect determining unit for determining the defective pixel address by evaluating a result of the plurality of times of detecting the candidate defect data based on a comprehensive standard for the plurality of times. Item 2. The imaging device according to Item 1. 3. The defect candidate data detecting means for detecting, as the black defect candidate pixel, the output of the image sensor in the normal exposure state when the level of the pixel is lower than a neighboring pixel by a predetermined level or more. The imaging device according to claim 2, comprising: 4. The defect candidate data detecting means, if the level of the pixel is smaller than a neighboring pixel by a predetermined level or more and substantially 0 with respect to the output of the image sensor in the normal exposure state, 3. The image pickup apparatus according to claim 2, further comprising: means for detecting as a pixel. 5. The defect candidate data detecting means for detecting, as the white defect candidate pixel, the output of the image sensor in the normal exposure state when the level of the pixel is higher than a neighboring pixel by a predetermined level or more. The imaging device according to any one of claims 2 to 4, comprising: 6. The defect candidate data detecting means, when the output of the image sensor in the normal exposure state is higher than a neighboring pixel by a predetermined level or more and is substantially saturated with a white defect, 5. A device according to claim 2, further comprising means for detecting a candidate pixel.
The imaging device according to any one of the preceding claims. 7. The defect determining means, wherein all of the plurality of times of defect candidate data detection executed by the defect candidate data detecting means are detected as the same type of defect candidate, and a pixel adjacent to the defect candidate pixel is detected. 3. The apparatus according to claim 2, wherein a pixel having a different level in the plurality of times is determined as a defective pixel.
The imaging device according to any one of claims 1 to 6. 8. When a registered defective pixel is present in the neighboring pixels, a pixel obtained by removing the registered defective pixel from the neighboring pixels is regarded as the neighboring pixel. The imaging device according to claim 3, wherein the defect data detection unit is configured to detect a defective pixel address. 9. The defect determining means is configured to determine, as a defective pixel, a pixel detected as a defect candidate of the same type in all of the plurality of times of defect candidate data detection executed by the defect candidate data detecting means. The imaging device according to any one of claims 2 to 6, wherein the imaging device is an image pickup device. 10. The photographing optical system is a variable-magnification optical system, and the defect data detecting means detects a plurality of defective pixel candidate addresses as a result of the plurality of defective pixel candidate addresses through a zoom operation of the photographing optical system. The imaging apparatus according to claim 2, wherein the detection of the defective pixel candidate address by the defect candidate data detection unit is performed using a result. 11. An image pickup apparatus capable of photographing a subject image using an image pickup device and recording image information obtained by the photographing, wherein a storage means for registering a defective pixel address of the image pickup device as defect data; Defect compensation means for performing a compensation process based on the defect data registered in the storage means with respect to the output from the image sensor in the above, and analyzing the output from the image sensor at each time of photographing. A defect detecting means for detecting a defective pixel candidate address and determining the defective pixel address based on a result of detecting the defective pixel candidate address a plurality of times.