JP3656711B2 - Image acquisition device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image acquisition apparatus for imaging a product from various angles using a plurality of television cameras, combining the plurality of images into one, and performing inspection by image processing.
[0002]
[Prior art]
Conventionally, as an apparatus for imaging an object using an imaging apparatus, a sensor, and a timer, there has been a video determination apparatus disclosed in Japanese Patent Laid-Open No. 9-298740. However, this conventional example can store the time taken by the sensor of the target object entering the sensor one after another using one TV camera, but using a plurality of TV cameras, Moreover, it was not possible to record in association with each same object.
[0003]
[Problems to be solved by the invention]
For example, when an object to be inspected is imaged from the entire periphery while being moved by a transport device such as a conveyor, the entire periphery cannot be imaged simultaneously because illumination is applied from multiple directions during imaging. In addition, due to the physical size of the TV camera, for example, the TV camera may not be arranged, and therefore, using a plurality of lighting devices and a plurality of TV cameras, the imaging position of each TV camera is shifted, Imaging is performed a plurality of times so that the illumination by the illumination device does not affect each other. For this reason, a sensor is provided at each imaging position, and the lighting device is activated at the moment when the product reaches a position suitable for imaging. When the sensor does not operate normally, an image that should have been captured cannot be obtained, or conversely, an unnecessary image is captured due to a malfunction of the sensor. If this occurs even once, there is a problem that the imaging results for the same product are shifted.
[0004]
Furthermore, if the time interval at which the product is introduced is not constant, a more complicated composition process is required because imaging cannot be performed at a constant interval. For this reason, when images are combined in the order of images captured by each television camera, images of different products may be erroneously combined.
[0005]
In addition, when image acquisition processing is performed by individually controlling the imaging timing of a plurality of television cameras by an image acquisition device having one CPU, if the number of television cameras increases, the control becomes complicated and the processing becomes complicated. The problem is that the speed is slow.
[0006]
The present invention has been made in view of the above problems, and images of products moving on a transfer device are picked up by a plurality of TV cameras from various angles, and the same product even if the input interval of the products is not constant. We propose a device that can acquire an image every time.
[0007]
[Means for Solving the Problems]
In order to achieve the above object, according to the first aspect of the present invention , a plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles, and depending on the position of the recognition object Means for determining the imaging timing of each imaging means; timing means for acquiring the imaging time of each imaging means; means for storing an image group obtained from each imaging means in association with the imaging time; an image acquisition apparatus you and means for collecting images of the same recognition target with a time difference, each time the same recognition target image group of acquisition is successful, update the information of time difference between the imaging time It has the means to perform, It is characterized by the above-mentioned.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
(Example 1)
FIG. 1 is a diagram showing an embodiment of an image acquisition apparatus according to the present invention. In the figure, 1-3 are television cameras, 11-13 are illumination devices, and 21-23 are sensors. The cylindrical product B is an inspection object, and a plurality of inspection objects are conveyed by the conveyor C in the direction indicated by the arrows. The feeding interval of each inspection object is not constant, but the conveying speed of the conveyor C is substantially constant.
[0009]
While the inspection object is transported by the conveyor, the inspection object is detected by the passage sensors 21 to 23, and using this as a trigger, the lighting devices 11 to 13, the timer, and the shutters of the television cameras 1 to 3 are operated for imaging. Signals from the TV cameras 1 to 3 are converted into digital images by a quantization device and stored in a storage device. At the same time, the time information obtained by the timer is stored in the storage device as data associated with the digital image. In addition, the image stored in the storage device is subjected to the combining process described in each of the following embodiments, and is stored in the storage device as a combined image.
[0010]
Now, when the passage of the product B is detected by the sensor 21, the time information of the timer (t = t1) is stored as the imaging time information by the television camera 1. Next, when the passage of the same product B is detected by the sensor 22, timer time information (t = t 2) is stored as imaging time information by the television camera 2. Assuming that the conveying speed of the conveyor C is V and the installation interval between the sensors 21 and 22 is L12, the time interval from when the same product B passes the sensor 21 to the next sensor 22 is t2−t1 = L12 / V. Next, when the product B further moves and the passage of the same product B is detected by the sensor 23, the time information (t = t3) of the timer is stored as the imaging time information by the television camera 3. If the conveying speed of the conveyor C is V, and the installation interval between the sensors 22 and 23 is L23, the time interval from the same product B passing through the sensor 22 to the next sensor 23 is t3−t2 = L23 / V.
[0011]
Therefore, it is determined that images having the same time interval are collected to be an image group of the same product. That is, in an inspection apparatus that inspects the product B from images captured using a plurality of television cameras 1, 2, 3, and in a configuration in which the input time interval of the product B is not constant, each of the television cameras 1, 2, 3 The imaging times t1, t2, and t3 are recorded, and those whose time intervals t2-t1 and t3-t2 are constant values are collected to determine that they are an image group of the same product. Time intervals t2-t1, t3-t2 are recorded when the first product passes. Alternatively, it is calculated in advance from the speed V of the conveyor C and the installation intervals L12, L23 of the sensors 21, 22, 23.
[0012]
Conditional expression 1 below is a conditional expression used to identify the same product in the configuration of three television cameras. (The number of TV cameras is 3 unless otherwise specified.)
Conditional expression 1:
(T2−t1) −Δt ≦ P12 ≦ (t2−t1) + Δt
And (t3−t2) −Δt ≦ P23 ≦ (t3−t2) + Δt
[0013]
However, t1, t2, and t3 are the imaging times of the television cameras 1, 2, and 3, respectively, P12 and P23 are parameters that indicate the difference in imaging time between the television cameras 1-2 and 2-3, and Δt is This is an error tolerance value for judging the same product.
[0014]
(Example 2)
In the first embodiment, the difference between the imaging time interval parameter and the time data gradually increases due to the time-dependent error of the timers of the television cameras 1, 2, and 3. If the timer errors of the television cameras 1 and 2 when 1 hour elapses are + T1 and −T2, respectively, 10 × (T1 + T2) is obtained after 10 hours elapses. Accordingly, every time the same product is identified, the difference in time data obtained at this time is reset as an imaging time interval parameter, thereby solving the problem of temporal error. In the above example, every time conditional expression 1 is satisfied, (t2-t1) is input to P12, and (t3-t2) is input to P23. As described above, by resetting the imaging time interval parameters P12 and P23 at the time of each photographing, it is possible to correct the time-dependent error of the timers attached to the television cameras 1, 2, and 3. Moreover, it can respond also to the case where the speed of the conveyor C fluctuates.
[0015]
(Example 3)
In the first embodiment, the value of the imaging time interval varies depending on the accuracy of the sensor of each television camera and the accuracy of the timer. Therefore, the difference in time data obtained each time the same product is identified is stored as an imaging time interval parameter a plurality of times, the average value of these values is calculated, and the imaging time interval parameters P12 and P23 are reset with the average value. By doing so, the problem of time accuracy can be solved. Assuming that N is an integer of 2 or more and the i-th time interval parameters from the newest when the images are taken at normal imaging intervals are P (i) 12 and P (i) 23, respectively.
P12 = ΣP (i) 12 / N
P23 = ΣP (i) 23 / N
It becomes.
[0016]
(Example 4)
FIG. 2 is an explanatory diagram of Embodiment 4 of the present invention. In this embodiment, when there is no sensor input for a relatively long time, such as when production is temporarily stopped, for example, when production has stopped and 10 minutes or more have elapsed, the imaging time interval parameters P12 and P23 are discarded. Next, the time when the first product passes when production is resumed is measured, and newly set as imaging time interval parameters P12 and P23. In the figure, the products shown in gray are the products that were first introduced when production was resumed. If the timing at which this product is detected and picked up by the sensors of the TV cameras 1, 2, and 3 is t = t1, t2, and t3, the product is reset as P12 = t2-t1 and P23 = t3-t2. This solves the problem of timer error over time when production is suspended. Further, it is possible to cope with a case where the speed of the conveyor C after the resumption of production is different from the previous one.
[0017]
(Example 5)
FIG. 3 is an explanatory diagram of Embodiment 5 of the present invention. The images picked up by the television cameras 1 to 3 are judged to be non-defective / defective using an inspection apparatus using image processing. As a result of this determination, the product determined to be defective must be discharged by the discharge device E. However, if the discharge signal is not accurately controlled, there is a problem that discharge cannot be performed or another product is discharged. Arise. Therefore, when the time from the imaging time (t = t3) to the discharge by the television camera 3 described above is T, the timer controlled by the sensor 23 is used for controlling the start timing of the discharge device. That is, the imaging time (t = t3) indicated by the timer when imaging the product determined to be defective by the inspection device (product shown in gray in the figure) is transferred to the ejection device, and the ejection device E is the product. Waits for a time to arrive at a predetermined position (t = t3 + T), after which a discharging operation is started and the product is discharged.
[0018]
(Example 6)
FIG. 4 is an explanatory diagram of Embodiment 6 of the present invention. Images captured by the TV cameras 1, 2, and 3 are transferred to an image memory, temporarily stored in a memory unit for short-term storage, and transferred to an image processing apparatus for inspection by image processing. . As a result of the inspection, if it is determined that the product imaged at the timing of t = t1, t2, t3 is defective, information on the imaging time is given to the memory unit as an image storage signal. The memory unit searches for an image at a timing: t = tx (x = 1, 2, 3) corresponding to the image storage signal given from the image processing device, and thus, among the image groups stored in the memory unit for a short period of time. Only the product image shown in gray in the figure is extracted and stored in a magnetic storage device such as a hard disk.
[0019]
For example, it is desirable to store only defective product images for a long period of time in order to consider later process improvements. However, although a storage device generally called a memory is high-speed, there are drawbacks such as that the storage is erased if the power is not supplied and the storage capacity is relatively small. In addition, what is generally called a magnetic storage device has features such as a semi-permanent storage and a large capacity, but has a drawback of a slow writing speed. Therefore, only the defective product image data is transferred to the magnetic storage device such as a hard disk from the image data that was on the memory at the inspection stage. Since the stored time parameter is specific to the image data at the time of imaging, the image data matching the time parameter is searched from the memory and transferred to the hard disk when transferring from the memory to the hard disk. To do.
[0020]
(Example 7)
FIG. 5 is an explanatory diagram of Embodiment 7 of the present invention. In this embodiment, a plurality of television cameras are simultaneously controlled by a trigger generated when the passage of a product is detected by a sensor, and images are taken from a plurality of directions at the same timing. For example, when the side and top surfaces of a product are imaged at the same time, and the side and top images are inspected by separate image processing, the side and top TV camera shutters are controlled based on the trigger signal from the sensor and imaged simultaneously. To do. In the illustrated example, first, the side camera 1a and the top camera 1b are simultaneously controlled by a trigger generated by one sensor, and a product is imaged from the side and top directions at the same timing. Next, the side camera 2a and the top camera 2b are simultaneously controlled by a trigger generated by another sensor, and the same product is imaged from the side direction and the top direction from different angles at the same timing. Images captured by the side cameras 1a and 2a are synthesized as side camera images and inspected by image processing. Further, the images taken by the top cameras 1b and 2b are synthesized as top camera images and inspected by another image processing.
[0021]
(Example 8)
FIG. 6 is an explanatory diagram of Embodiment 8 of the present invention. The present embodiment is characterized in that, with respect to an image of a television camera that images a product, a steady shift amount (parallel movement, enlargement / reduction, rotation, etc.) at the time of imaging is corrected by means such as normalization. . Ideally, multiple TV cameras that capture images of the product should maintain constant conditions such as position in the field of view, magnification, and rotation (tilt), but in reality, the position of the TV camera cannot be completely adjusted. Therefore, a steady shift occurs in the obtained image. Therefore, after the correction calculation is performed on the obtained image, it is stored in the storage device. In the illustrated example, when adjacent TV cameras 1 and 2 are capturing images that are regularly translated, parallel processing is performed on one or both of the data transferred from the TV cameras 1 and 2. Are stored in the storage device.
[0022]
Example 9
FIG. 7 is an explanatory diagram of Embodiment 9 of the present invention. In this embodiment, a virtual development view is created on an image memory by copying image data as necessary in consideration of the positional relationship of imaging by a TV camera, and a wider and more accurate inspection is performed. is there. In the figure, numerals 1 to 3 indicate the field of view of each television camera (three in this example). By copying the field of view of the television camera 1 on the image memory, an inspection on the boundary between the television camera 3 and the television camera 1 can be performed.
[0023]
FIG. 8 is a top view of a product to be inspected according to this embodiment. In the illustrated example, the entire circumference of a cylindrical product is inspected for defects by synthesizing three images obtained by imaging from three different angles with three television cameras. When a defect exists between 1 and the visual field of the television camera 3, the defect image is divided, so that the defect detection accuracy by the inspection apparatus using the subsequent image processing is lowered. Therefore, based on the geometric positional relationship, the image of the television camera 1 is written twice in the image memory as shown in FIG. By doing in this way, it becomes possible to process continuously the image of the boundary part of the visual field of the television camera 3 and the television camera 1.
[0024]
【The invention's effect】
According to the first to seventh aspects of the present invention, the same recognition target image group is collected with the time difference of the imaging time from the image group captured from a plurality of different angles at different timings by the plurality of imaging means. If the images are conveyed at a substantially constant speed, a group of images can be collected for each identical recognition target even if the input intervals are irregular.
According to the first to third aspects of the present invention, the time-dependent error of the time measuring means can be eliminated by appropriately updating the information on the time difference of the imaging time, and it is also possible to cope with the fluctuation of the conveyance speed.
[0025]
According to the invention of claim 4 , since the secondary storage device for storing the collected image group of the same recognition target is provided, for example, only the defective product image is taken for a long time for the examination of the later process improvement and the like. It can be saved.
[0026]
According to the fifth aspect of the present invention, since the plurality of imaging units are operated at the same imaging timing corresponding to one position of the recognition target and the recognition target is simultaneously imaged from a plurality of different angles, The number of sensors that detect the position can be saved.
According to the sixth aspect of the present invention, since the image group obtained from the plurality of image pickup means is provided with the means for correcting the shift amount between the image pickup means, the positioning accuracy of the image pickup means may be low, and at the time of installation The adjustment work becomes easier.
According to the seventh aspect of the present invention, it is possible to continuously process the image of the boundary portion of the field of view of the image pickup means, so that it is possible to prevent the defect detection accuracy from deteriorating depending on the location where the defect exists.
[Brief description of the drawings]
FIG. 1 is a perspective view showing an embodiment of an image acquisition apparatus of the present invention.
FIG. 2 is an explanatory diagram of Embodiment 4 of the present invention.
FIG. 3 is an explanatory diagram of Embodiment 5 of the present invention.
FIG. 4 is an explanatory diagram of Embodiment 6 of the present invention.
FIG. 5 is an explanatory diagram of Embodiment 7 of the present invention.
FIG. 6 is an explanatory diagram of Embodiment 8 of the present invention.
FIG. 7 is an explanatory diagram of Embodiment 9 of the present invention.
FIG. 8 is a plan view of a product imaged according to Embodiment 9 of the present invention as viewed from above.
[Explanation of symbols]
1, 2, 3 TV camera 11, 12, 13 Illuminator 21, 22, 23 Sensor C Conveyor B Product

Claims (7)

A plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles;
Means for determining the imaging timing of each imaging means according to the position of the recognition target;
Time measuring means for obtaining the imaging time of each imaging means;
Means for storing the image group obtained from each imaging means in association with the imaging time;
With a time difference of the imaging time by an image acquisition device you and means for collecting images of the same object of recognition,
An image acquisition apparatus comprising: means for updating information on a time difference between imaging times each time an image group of the same recognition target is normally collected . Means for recording time difference information of the imaging time each time collection of the same recognition target image group is normally performed, and an imaging value obtained by averaging the time difference information of the recorded imaging times for the latest multiple times image acquisition apparatus according to claim 1, characterized by comprising means for setting the information of the time difference. If a recognition target is not detected for a certain time or longer, a unit for discarding the time difference information of the imaging time and resetting the information of the time difference of the imaging time based on the imaging time measured for the next input product is provided. The image acquisition apparatus according to claim 1 . A plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles;
Means for determining the imaging timing of each imaging means according to the position of the recognition target;
Time measuring means for obtaining the imaging time of each imaging means;
Means for storing the image group obtained from each imaging means in association with the imaging time;
Means for collecting the same recognition target image group with a time difference in imaging time,
Of the same image group to be recognized, which is collected based on the time difference of the imaging time, the image acquiring apparatus characterized by comprising a secondary storage device for storing the images for a specific recognition target. A plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles;
Means for determining the imaging timing of each imaging means according to the position of the recognition target;
Time measuring means for obtaining the imaging time of each imaging means;
Means for storing the image group obtained from each imaging means in association with the imaging time;
Means for collecting the same recognition target image group with a time difference in imaging time,
The means for determining the imaging timing of each imaging means according to the position of the recognition target operates a plurality of imaging means at the same imaging timing corresponding to one position of the recognition target, and simultaneously selects the recognition target from a plurality of different angles. image capture device characterized by imaging. A plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles;
Means for determining the imaging timing of each imaging means according to the position of the recognition target;
Time measuring means for obtaining the imaging time of each imaging means;
Means for storing the image group obtained from each imaging means in association with the imaging time;
Means for collecting the same recognition target image group with a time difference in imaging time,
For images obtained from a plurality of image pickup means, the image capture device characterized by comprising a means for correcting the amount of deviation between the image pickup means. A plurality of imaging means for imaging a recognition object conveyed at a substantially constant speed from a plurality of different angles;
Means for determining the imaging timing of each imaging means according to the position of the recognition target;
Time measuring means for obtaining the imaging time of each imaging means;
Means for storing the image group obtained from each imaging means in association with the imaging time;
Means for collecting the same recognition target image group with a time difference in imaging time,
A plurality of different angles the same image group of the recognition target captured from the image acquisition device you characterized by comprising a means for storing as expanded image based on the geometric positional relationship between the recognition target.

Images