JP3869502B2 - Imaging method and CCD camera - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an imaging method and a CCD camera used when image processing is performed using natural light such as machine vision in FA (automated factory) such as visual inspection of electronic parts and automatic monitoring of traffic congestion on a road. More specifically, the present invention relates to a wide dynamic imaging method and a CCD camera that can obtain a clear image as a whole even if the brightness varies depending on the location (part) of the subject.
[0002]
[Prior art]
When performing an appearance inspection of parts or products by image recognition in machine vision or the like in FA, for example, near a window or near an entrance, the brightness of a subject varies depending on the location due to incident light from the sun. Also, at the entrance and exit of the tunnel, a bright part (high brightness subject) and a dark part (low brightness subject) are mixed. In such a case, as shown in FIG. 5A, the bright image A is displayed whitish and the dark image C becomes dark, and even if the shutter speed is adjusted, the image recognized by the CCD camera is clear as a whole. Cannot be obtained and accurate image processing cannot be performed. B indicates that the brightness is appropriate.
[0003]
In order to solve such a problem, for example, Yamada et al., “Positioning of parts resistant to lighting fluctuations by wide dynamic range vision” (7th “Appearance Inspection” held at Kogakuin University on November 1-2, 1995) (Automation Workshop, pages 27-32) discloses a method of using a shutter camera and taking two or more images of the same object with different shutter speeds and recombining them. That is, two images, an image P having a slow shutter speed as shown in FIG. 5B and an image Q having a fast shutter speed as shown in FIG. A method of synthesizing parts.
[0004]
In addition, the E-1 “Wide Dynamic Range Camera” (issued June 1996, pp. 151 to 154), the second image sensing symposium sponsored by the Image Sensing Technology Study Group, has a non-destructive readout function, and exposure time A method is disclosed in which two different images are read out at a field rate and combined to obtain a linear video signal. That is, as shown in FIG. 6, a signal of a short exposure time drawn nondestructively in the first 16H time in one field period (1/60 seconds) from each photosensor and at the end of one field period. Data with a long exposure time is transferred to the signal processing unit and synthesized.
[0005]
[Problems to be solved by the invention]
The above-described method of synthesizing images with different shutter speeds is not suitable for recognizing a moving object because there is a time difference because the frames from which the images are obtained differ. Furthermore, it takes two or more frames to create one screen, and there is a problem that it takes a long time to obtain a final image.
[0006]
In the above-described method of reading two images with different exposure times, since the data is in the same frame, the time lag is small, but an image with a field period of, for example, 1/60 seconds and a short exposure time is Since the exposure is performed at about 1/16 of the first time, for an image exposed in the entire time and an image exposed in a short time, the maximum time is 1/60 × 15/16 = 0 in the time for recognizing the image. A time lag of about .016 seconds occurs. Therefore, in high-speed imaging such as an automobile, image shift occurs and the sense of incongruity cannot be wiped off.
[0007]
The present invention solves such a problem, and an image of a subject in which a bright portion (high luminance portion) and a dark portion (low luminance portion) are mixed in a high-speed moving subject is optimally exposed and has less blur. An object is to provide an imaging method and a CCD camera that can obtain an image.
[0008]
[Means for Solving the Problems]
Imaging method of the present invention is an imaging method for obtaining an image signal by transferring charges produced by received by each photosensor arranged in a matrix, wherein each photosensor or found 3 times charge during one field period the drawer, to create a sum of the signals by the first and last drawer, and one entire field signal, to employ either signal a signal of a normal person is compared with a reference signal of the two signals Rukoto Thus, the exposure time varies depending on the pixel. As a result, a signal with a short exposure time can be brought to the average middle part of the long exposure time, and the sense of incongruity between the two is reduced.
[0009]
Another embodiment of the imaging method according to the present invention is an imaging method for obtaining a video signal by transferring charges generated by receiving light by each photosensor arranged in a matrix, from each photosensor during one field period. The fourth charge is extracted, the second charge is extracted at the center of one field period, and the sum of the signal from the first extraction, the signal from the third extraction, and the signal of the entire field are created. By adopting the normal signal by comparing one of the two signals with the reference signal, the exposure time can be made different depending on the pixel. The second charge extraction is performed so that the second charge extraction becomes data in a period sandwiching the middle part of the one field period. Create a Rudo overall signal, by one of the signals of the two signals is compared with a reference signal to adopt a normal person signal, it is also possible to vary the exposure time by pixels.
[0010]
CCD camera according to the present invention, the photosensor and the sensor group arranged in a matrix, and a charge transfer section to draw out the charge is divided into at least three times during one field period from the photosensor, and transfers the signal processing unit, The signal processing unit extracts at least three times and is a signal based on the sum of the charges extracted on both sides of the intermediate part of the one field period or a signal based on the entire one field period and a reference signal a comparator for comparing, that has a switching circuit for selecting either of a signal by the whole or signal by the sum by the output of the comparator.
[0011]
A sensor group in which photosensors are arranged in a matrix; a charge transfer unit that extracts charges from each photosensor at least three times during one field period and transfers the charges to a signal processing unit; and the signal processing unit, A comparator for comparing a reference signal with a signal derived from electric charges drawn at least three times and drawn during a time spanning an intermediate portion of the one field period and a signal derived from the entire one field period; and the comparator And a switching circuit that selects either a signal based on charges drawn during the time spanning the intermediate portion or a signal based on the whole .
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Next, an imaging method and a CCD camera of the present invention will be described with reference to the drawings.
[0013]
FIG. 1 is a diagram showing an embodiment of a CCD camera of the present invention. In FIG. 1, reference numeral 11 denotes a photosensor made of, for example, a photodiode, which is arranged in a matrix to constitute each pixel. Reference numerals 12a, 12b and 12c are transfer gates schematically showing that the charge of the signal received by the photosensor 11 is extracted after a predetermined time t ₁ , t ₂ and t ₃ in one field. The extracted charge is tripled. The charge transfer unit 14 in the vertical direction and the charge transfer unit 15 in the horizontal direction are transferred at a speed of 1 to the signal processing unit 20. That is, at least 3 times the charge is issued can pull in the present invention from one photosensor 11 constituting one pixel in one field period, and resetting the charge accumulated during the last drawer, at the next field It is what starts charge accumulation.
[0014]
The charge extraction timing will be described with reference to FIG. FIG. 2 is a diagram showing a change in the amount of charge accumulated in one field period by one photosensor 11. The horizontal axis is the time of one field, for example, 1/60 seconds, and the vertical axis is the charge accumulated in the photosensor 11. Indicates the amount. In FIG. 2, E, F, and G are times for drawing out charges from the above-described photosensor 11, and the above-mentioned fixed times t ₁ , t ₂ , and t ₃ indicate the times of OE, OF, and OG, respectively. The time t ₁ is determined by the difference in luminance between the bright (high luminance) portion and the dark (low luminance) portion of the subject, and is determined by a fraction of the dark portion of the subject. When the exposure time for the bright part is set to 1/16 of the total time for exposing the dark part (time for one field), the exposure time is set to be half that time.
[0015]
Finally, the time t ₃ for extracting the charge is the last time of one field, but the time t ₂ for extracting the second charge is t ₃ -t ₁ in the example shown in FIG. That is, in the present invention, the exposure time of the bright portion is not exposed at the beginning of the field, but is exposed separately in half at the beginning and the end of the field period, and the amount of charge due to both exposures is added up. To do. As a result, even if the subject moves in one field, it is recognized as the average position of the first and last in one field, and the positional deviation between the bright part and the dark part of the subject can be reduced.
[0016]
The signal processing unit 20, charges accumulated in each time is transferred to the first signal line 21 as a signal of a dark object portion is the charge drawn at time t _3, t ₃ calculated by the calculation unit 29 The charges extracted and transferred at the time of -t ₂ + t ₁ are sent to the second signal line 22 as a signal of the bright subject portion, and are amplified by the amplifiers 23a and 23b, respectively. Then, the sample-and-hold circuit 24a, the signal drawn at each time t ₃ at 24b and t ₃ -t by the clock signal from the clock signal generator 28 to a clock signal by dividing for transferring charges in the horizontal direction charge transfer section 15 _The signal extracted at time ₂ + t ₁ is separated. Among the separated signals, for example, a part of the signal from the first signal line 21 is branched and compared with the reference signal by the comparator 25 to determine whether or not the transferred signal has normal brightness. If the signal compared by the comparator 25 is normal, the signal on the first signal line 21 is selected by the switching circuit 27 composed of a multiplexer, for example, and the signal on the second signal line 22 is selected and output. Note that the comparator 25 may compare with the signal of the second signal line 22.
[0017]
The signal of the second signal line 22 is amplified by the calibration amplifier 26 and returned to a state where the sensitivity is not lowered, and the level is matched with the signal of the first signal line. This calibration amplifier 26 does not need to be used if the shortening of the exposure time known in advance is corrected, but can be adjusted accurately by making it programmable by the amplifier.
[0018]
With the above configuration, signal levels can be compared for each pixel provided with a sensor, and a clear image signal can be selected and used as image data. A clear image as a whole can be obtained even when a subject with too much part is photographed. In addition, the exposure time for the bright part is divided during the field period, and the charge amount obtained by each exposure is added, so even if the subject is moving, an intermediate period in one field period is obtained. It is recognized as a position, and there is no sense of incongruity with the dark image signal. For example, when there is a bright part L and a dark part D as shown in FIG. 3A and an object moving with a width W is imaged, as shown in FIG. 3B, the dark part D is a field period. Since the middle image is taken, the image looks bright at the part where the moving objects overlap in one field period, and the two ends are blurred and dark (the dark part is shaded). Can be recognized. On the other hand, the bright portion L is the beginning and end of one field period (t ₃ ), and time images of t ₁ and t ₃ -t ₂ respectively appear, and these images are similarly bright at the center and at both ends. Although it is darkly blurred, the subject is originally bright and the exposure time is short, so there are few areas that are blurred. The interval between the two images differs depending on the width W of the subject and the relative movement speed of the subject, and may overlap each other. If the two images overlap, they are recognized as one unit within the field period, but even if they do not overlap, the images can be recognized again in a short time and appear as one unit. As a result, it is recognized as an integral image in which the brightness is matched with the dark subject portion. On the other hand, in the conventional two-division method in which the bright part is imaged for the time t ₅ only by the first part of one field period, as shown in FIG. 3C, the dark part is the entire field period. On the other hand, since the bright part is recognized only at the beginning of the field period, there is a sense of incongruity that the shape of the bright part appears flying even if the brightness matches. In the present invention, such a sense of incongruity can be eliminated.
[0019]
In the example shown in FIG. 1, performs a drawer of charge in t ₁ and t _2, using the sum of t ₁ of the charge and t ₃ of the charge when a short exposure time, the t ₁ when the long exposure time charges and and using the sum of the charge of the charge and t ₃ of t _2.
[0020]
FIG. 4 is a diagram showing another example of the timing of charge extraction . In FIG. 4B , t ₂ -t ₁ becomes an exposure time to be shortened (in the above example, 1/16 of the field period (for example, 1/60 second)) in the middle portion M of one field period. t ₁ and t ₂ are set, and t ₃ is set at the end of the field period. By doing so, the center times of the imaging data of the bright subject portion having a short exposure time and the dark subject portion having a long exposure time are equalized, and the sense of incongruity between the two is reduced.
[0021]
4 (a) is an example of performing four times the drawer charge. That is, in the example shown in FIG. 2, the time for extracting the times t ₂ and t ₃ is brought to the center M of one field period, and the charge at the end of one field period is extracted at t ₄ . As described above, by performing the charge extraction four times, the imaging data of the bright subject portion can be used as the sum of the data of an arbitrary portion. In this case, four CCDs are provided and read out at a quadruple speed rate. In addition to the above-described example, the number of charge transfer units (14 in FIG. 1) by CCD can be increased instead of increasing the reading rate. Further, by further increasing the number of charge extractions from four, data with a short exposure time can be further subdivided, and the sense of incongruity with data with a long exposure time can be further alleviated.
[0022]
【The invention's effect】
According to the present invention, since the charge extraction from each photosensor is performed three times or more, the short exposure time data can be obtained as the data of the central portion of one field period or the sum of data of two or more locations. it can. For this reason, an image with a wide dynamic range with little discomfort can be obtained even with an image of a subject moving at high speed with little difference from an image based on data of a long exposure time.
[0023]
As a result, a clear image can be obtained even when the vehicle is monitored at the entrance of the tunnel or when a moving object is imaged in the backlight state, and can be effectively used for automatic monitoring of the vehicle.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of an embodiment of a CCD camera of the present invention.
FIG. 2 is a diagram for explaining a charge extraction timing of the CCD camera of the present invention.
FIG. 3 is a diagram illustrating how images of a bright subject portion and a dark subject portion appear by the CCD camera of the present invention.
FIG. 4 is an explanatory diagram of another example of charge extraction timing of the CCD camera of the present invention.
FIG. 5 is an explanatory diagram of a conventional image of subjects with different brightness.
FIG. 6 is an explanatory diagram of charge extraction timing in an example of obtaining a short exposure time and a long exposure time by two conventional charge extractions.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 Photosensor 14 Charge transfer part 15 Charge transfer part 20 Signal processing part 25 Comparator 27 Switching circuit 29 Calculation part

Claims (5)

An imaging method for obtaining an image signal by transferring charges produced by received by each photosensor arranged in a matrix form, drawn out the respective photosensor or found 3 times charge during one field period, the first and last the sum of the signals by the drawer, to create a one-field total signal by Rukoto to employ either signal a signal of a normal person is compared with a reference signal of the two signals, the exposure time by pixels Different imaging methods. An imaging method for obtaining a video signal by transferring charges generated by receiving light by each photosensor arranged in a matrix, wherein the charges are extracted four times from each of the photosensors during one field period. The drawer is performed at the center of one field period, and the sum of the signal from the first drawer and the signal from the third drawer and the signal of the entire field are created, and either of the two signals is used as a reference signal. An imaging method in which the exposure time differs depending on the pixels by adopting the signal that is normal compared to . An imaging method for obtaining a video signal by transferring charges generated by receiving light received by each photosensor arranged in a matrix, wherein the charges are extracted three times from each of the photosensors during one field period. The extraction is performed so that the data is in a period sandwiching the middle part of one field period, and a signal from the second extraction and a signal for the entire one field are generated, and one of the two signals is used as a reference signal. An imaging method in which the exposure time varies depending on the pixels by adopting the signal of the normal one in comparison . Photosensor and the group sensors arranged in a matrix, and pull out the charge is divided into at least three times during one field period from the photosensor, a charge transfer unit that transfers the signal processing section, by the signal processing unit, wherein A comparator for comparing a reference signal with a signal derived from at least three times and extracted from both sides of the middle portion of the one field period and a signal of one of the signals of the entire one field period and the reference signal; CCD camera ing and a switching circuit for selecting either of a signal by the whole or signal by the sum by the comparator output. A group of sensors in which photosensors are arranged in a matrix, a charge transfer unit that extracts charges from each photosensor at least three times during one field period, and transfers them to a signal processing unit. A comparator for comparing a reference signal with a signal derived from electric charges drawn at a time that extends in the middle of the one field period and a signal derived from the entire one field period, and an output of the comparator switching circuit and a CCD camera ing a for selecting one of the signal by the whole or a signal due to charge drawn in time across the intermediate portion by.

Images