JP6874671B2 - Authentication system and data processing method - Google Patents

Description

The present invention relates to an authentication system and a data processing method for performing object fingerprint authentication.

Conventionally, as shown in Patent Documents 1 to 5 and Non-Patent Document 1, object fingerprint authentication is known.

Object fingerprint authentication realizes individual identification (authentication) by imaging an individual-specific surface pattern (object fingerprint) that occurs spontaneously in the manufacturing process of a product or part as an image and collating the images. According to the object fingerprint authentication, even a minute part or a part that cannot be processed on the surface can be identified by simply taking a magnified image with a camera under appropriate lighting conditions.

Japanese Unexamined Patent Publication No. 2016-100527 Japanese Unexamined Patent Publication No. 2016-176753 Japanese Unexamined Patent Publication No. 2016-178158 Japanese Unexamined Patent Publication No. 2017-058306 Tokukai 2017-07936

Rui Ishiyama, 2 outsiders, "Individual identification by object fingerprint authentication that expands the use of IoT", [online], NEC Technical Report Vol.69 (2016) No.1 (September), [November 28, 2017 Day search], Internet <http://jpn.nec.com/techrep/journal/g16/n01/160110.html>

From the viewpoint of traceability, there is a demand for individual identification of an object such as a work in the middle or final stage of various processes such as a production process. When such identification is performed by object fingerprint authentication, it is necessary to image the object fingerprint at a stage prior to these stages and pre-register the feature amount data of the object fingerprint in a device such as an image processing device. There is.

In this case, the number of feature data registered in the device increases with the passage of time. When the number of feature data increases in this way, the possibility of erroneous authentication increases. Further, as the number of feature data increases, the number of matchings increases, so that the time required for authentication increases.

The present invention has been made in view of the above problems, and an object of the present invention is to provide an authentication system and a data processing method capable of suppressing false authentication and executing authentication in a short time.

According to certain aspects of the invention, the authentication system captures a first portion of a plurality of objects, each transported along a predetermined path, at a first position upstream of the path. A camera, a second camera that images a first part of each object at a second position on the downstream side of the path, and an image processing device communicatively connected to the first camera and the second camera. To be equipped with. The image processing device is a generation means for generating first feature amount data representing the feature amount of the object fingerprint of the first part for each object based on the image data of the first part acquired from the first camera. And a storage means for storing each of the generated first feature amount data. The generation means further generates second feature amount data representing the feature amount of the object fingerprint of the first part for each object based on the image data of the first part acquired from the second camera. The image processing apparatus includes a determination means for determining whether or not the first feature amount data matching the second feature amount data exists in the storage means based on the generation of the second feature amount data, and a second. It further includes an erasing means for erasing the first feature amount data that matches the feature amount data of 2 from the storage means.

According to the above configuration, false authentication can be suppressed and authentication can be executed in a short time.
Preferably, the authentication system captures a second portion of each object at a third position upstream of the path and a second portion of each object at a fourth location downstream of the path. It is further provided with a fourth camera that captures images at the position of. The generation means further generates a third feature amount data representing the feature amount of the object fingerprint of the second part for each object based on the image data of the second part acquired from the third camera. Based on the image data of the second part acquired from the fourth camera, the fourth feature amount data representing the feature amount of the object fingerprint of the second part is generated for each object. The storage means stores the first feature amount data and the third feature amount data in association with each object. The determination means further determines whether or not the third feature amount data that matches the fourth feature amount data exists in the storage means based on the generation of the fourth feature amount data. The erasing means is a case where the first feature amount data matching the second feature amount data exists in the storage means and the third feature amount data matching the fourth feature amount data exists in the storage means. , The first feature amount data and the third feature amount data are deleted from the storage means on condition that the first feature amount data and the third feature amount data are associated with each other.

According to the above configuration, since the image processing apparatus performs authentication using feature amount data of a plurality of parts, it is possible to improve the authentication accuracy.

Preferably, the second camera takes a plurality of imagings for each object. The generation means generates a plurality of second feature amount data representing the feature amount of the object fingerprint of the first portion for each object based on a plurality of times of imaging. The determination means is based on the fact that a plurality of second feature amount data have been generated, and the first feature amount data that matches a predetermined number or more of the second feature amount data among the plurality of second feature amount data. It is determined whether or not the feature data of the above is present in the storage means. The erasing means erases the first feature amount data that matches the predetermined number or more of the second feature amount data from the storage means.

According to the above configuration, since the image processing device performs authentication using a plurality of feature amount data, it is possible to improve the authentication accuracy.

Preferably, the determination means has a first feature that matches the second feature data, provided that the similarity between the second feature data and the first feature data is equal to or greater than a predetermined threshold value. It is determined that the feature data exists in the storage means. The image processing apparatus further includes adjusting means for adjusting the threshold value. When there are a plurality of first feature amount data matching the second feature amount data in the storage means, the adjusting means has one first feature amount data that matches the second feature amount data. To raise the threshold.

According to the above configuration, the threshold Th can be set to an appropriate value.
Preferably, the adjusting means has one first feature amount data that matches the second feature amount data when the first feature amount data that matches the second feature amount data does not exist in the storage means. Lower the threshold until it is determined to exist.

According to the above configuration, the threshold Th can be set to an appropriate value.
Preferably, when the first feature amount data that matches the second feature amount data does not exist in the storage means, the adjusting means lowers the threshold value up to a predetermined lower limit value.

According to the above configuration, it is possible to prevent the value of the threshold value Th from becoming equal to or less than the lower limit value.
Preferably, the authentication system further comprises a display device communicatively connected to the image processor. In the image processing device, when a plurality of first feature amount data matching the second feature amount data exists in the storage means, or when the first feature amount data matching the second feature amount data exists in the storage means. If not, the display device is made to perform a predetermined display.

According to the above configuration, the user can set the threshold value to an appropriate value.
According to another aspect of the invention, the data processing method images a predetermined portion of a plurality of objects, each transported along a predetermined path, at a first position on the upstream side of the path. Communicatably connected to a first camera, a second camera that captures a predetermined portion of each object at a second position downstream of the path, a first camera, and a second camera. It is performed in an authentication system equipped with an image processing device. The data processing method generates first feature amount data representing the feature amount of the object fingerprint of the predetermined part for each object based on the image data of the predetermined part acquired from the first camera. Based on the step, the step of storing the generated first feature amount data in the memory, and the image data of the predetermined part acquired from the second camera, the feature amount of the object fingerprint of the predetermined part. Based on the step of generating the second feature amount data representing the above for each object and the generation of the second feature amount data, the first feature amount data that matches the second feature amount data is It includes a step of determining whether or not it exists in the memory, and a step of erasing the first feature amount data that matches the second feature amount data from the memory.

According to the above method, false authentication can be suppressed and authentication can be executed in a short time.

According to the present invention, it is possible to suppress erroneous authentication and execute authentication in a short time.

It is a figure which showed the schematic structure of the authentication system. It is a figure which showed the specific example of the authentication system. It is a functional block diagram which showed the functional structure of an image processing apparatus. It is a block diagram which showed the hardware included in an image processing apparatus. It is a figure showing the fingerprint of the object in a part area F1 of the side surface of a work. It is a figure showing the fingerprint of the object in a part area F2 of the surface of a work. The correlation between the feature amount data of the pre-registered model board and the feature amount data of the other three boards and the feature amount data separately acquired for the model board when the authentication is performed using the object fingerprint on the side surface is shown. It is a figure. The correlation between the feature amount data of the pre-registered model board and the feature amount data of the other three boards and the feature amount data separately acquired for the model board when the authentication is performed using the object fingerprint on the surface is shown. It is a figure. It is a flow diagram for demonstrating the process flow of the pre-processing. It is a figure which showed the flow of the process until the feature amount data is registered in a memory. It is a figure which showed the flow of the process until the feature data is erased from a memory. It is a figure which showed the transition of the number of feature quantity data stored in a memory in the case where the data amount was adjusted by the erasing process, and the case where the data amount was not adjusted. It is a figure which showed the flow of the process until the feature data is erased from a memory. It is a flow chart which showed the flow of the threshold value setting process. It is a figure which showed the specific example of another authentication system. It is a figure for demonstrating the position of two parts. It is a figure which showed the flow of the process until the feature amount data about two parts is registered in a memory. It is a figure which showed the flow of the process until the feature data is erased from a memory.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are designated by the same reference numerals. Their names and functions are the same. Therefore, detailed explanations of them will not be repeated.

§1 Application example FIG. 1 is a diagram showing a schematic configuration of the authentication system 1.

With reference to FIG. 1, the authentication system 1 includes a camera 10, a camera 20, and an image processing device 30. The camera 10 and the camera 20 are communicably connected to the image processing device 30.

The camera 10 images a predetermined portion (hereinafter, referred to as “site B1”) of a plurality of objects, each of which is conveyed along a predetermined path, at a position Pa on the upstream side of the path. The camera 20 images the portion B1 of these objects at the position Pb on the downstream side of the path.

In the example of FIG. 1, the cameras 10 and 20 take an image of the portion B1 of the work W, which is an object conveyed on the path by the transfer device 900. The cameras 10 and 20 transmit the image data obtained by imaging to the image processing device 30. The work W may be extracted on the route.

The image processing device 30 generates feature amount data R representing the feature amount of the object fingerprint of the part B1 for each work W based on the image data of the part B1 acquired from the camera 10. The image processing device 30 stores each generated feature amount data R in the memory. In the image processing device 30, typically, an upper limit value of the number of feature amount data R stored in the memory is set. The upper limit is typically determined in advance based on the number of workpieces flowing through the path. In this example, the upper limit is set to 100.

In the case of the example of FIG. 1, as the state (a), 100 feature amount data R # 1 to # 100 are registered as registered data in the memory (specifically, the database in the memory).

The image processing device 30 further generates feature amount data T representing the feature amount of the object fingerprint of the part B1 for each work W based on the image data of the part B1 acquired from the camera 20.

Based on the fact that the feature amount data T is generated, the image processing device 30 determines whether or not the feature amount data R that matches the feature amount data T exists in the memory. For example, when the state of the memory is the state (a) at the time when the feature amount data T # 1 is generated, the image processing apparatus 30 has the feature amount data T # 1 and 100 feature amount data R # 1. Collate with ~ # 100.

The image processing device 30 erases the feature amount data R that matches the feature amount data T from the memory. In the case of the example of FIG. 1, when the feature amount data T # 1 and the feature amount data # R1 match, the image processing apparatus 30 deletes the feature amount data R # 1 from the database as shown in the state (b). To do. After that, based on the imaging of the work W by the camera 10, the feature amount data # 101 is additionally registered in the database as shown in the state (c).

After that, the above-mentioned determination process (verification process) and the above-mentioned erasure process are repeated. For example, after the feature amount data R # 1 is erased, the image processing apparatus 30 combines the feature amount data T # 2 with the 100 feature amount data # 2 to # 101 in the database shown in the state (c). Perform matching. For example, when the feature amount data T # 2 and the feature amount data # R2 match, the image processing apparatus 30 deletes the feature amount data R # 2 from the database.

In the above, the configuration for erasing the data in the database is given as an example, but the configuration may be such that the data is stored in a file, on-memory, or the like instead of the database. The storage location of the feature data R is not particularly limited.

According to such a configuration, the feature amount data R stored in the memory is sequentially erased on condition that the authentication is successful. As a result, it is possible to suppress an increase in the number of feature data R stored in the memory. Therefore, false authentication can be suppressed and authentication can be performed in a short time.

In the case of the above configuration example, the upper limit value may be set to a value larger than 100 (for example, 120) in consideration of the existence of the feature amount data R that remains without being erased due to the failure of authentication. preferable.

§2 Configuration example [Embodiment 1]
<A. System configuration>
FIG. 2 is a diagram showing a specific example of the authentication system 1.

With reference to FIG. 2, the authentication system 1 includes cameras 10 and 20, an image processing device 30, a PLC (Programmable Logic Controller) 40, a server device 50, and a display 60.

As described above, the image processing device 30 is communicably connected to the cameras 10 and 20. The image processing device 30 is communicably connected to the PLC 40. The PLC 40 is communicably connected to the server device 50 as a higher-level device. Further, the image processing device 30 causes the display 60 to display a predetermined image. In such a system configuration, the image processing device 30 performs object fingerprint authentication based on the image data acquired from the cameras 10 and 20.

The authentication result is displayed, for example, on the display 60 connected to the image processing device 30. Further, the authentication result may be transmitted to the server device 50 via the PLC 40. In this case, the user of the server device 50 can visually recognize the authentication result also in the server device 50.

The image processing device 30 is typically a visual sensor. Further, the work W is photographed by the camera 10 installed at the position Pa at the position P1, and is photographed by the camera 20 installed at the position Pb at the position P2.

<B. Image processing device>
FIG. 3 is a functional block diagram showing a functional configuration of the image processing device 30.

With reference to FIG. 3, the image processing device 30 includes an acquisition unit 31, an authentication processing unit 32, a storage unit 33, and an erasing unit 34. The authentication processing unit 32 includes a generation unit 321, a determination unit 322, and an adjustment unit 323.

The acquisition unit 31 acquires image data from each of the camera 10 and the camera 20.
The authentication processing unit 32 performs authentication (verification) using an image matching method using a local feature amount and an algorithm for verifying the consistency of the geometric arrangement of feature points.

The generation unit 321 generates feature data R representing the feature amount of the object fingerprint of the part B1 for each work W based on the image data of the part B1 acquired from the camera 10. The generation unit 321 further stores each generated feature amount data R in the storage unit 33. Further, the generation unit 321 generates feature amount data T representing the feature amount of the object fingerprint of the part B1 for each work W based on the image data of the part B1 acquired from the camera 20.

Specifically, the generation unit 321 first determines, as a feature point, a place where the change in brightness is steep and the position can be stably obtained from the fingerprint image of the object. Next, the generation unit 321 converts the local luminance pattern around the feature point into data as a feature amount. Through these processes, the generation unit 321 generates feature data R and T.

The determination unit 322 determines whether or not the feature amount data R that matches the feature amount data T exists in the storage unit 33 based on the generation of the feature amount data T. Specifically, the determination unit 322 executes the following processing.

The determination unit 322 obtains a pair of feature points having the smallest difference in feature amounts from both images to be collated (specifically, feature amount data R and T). The determination unit 322 verifies the geometrical consistency of the overall arrangement of the feature points. The determination unit 322 extracts only the pair group whose relative positional relationship with other feature points is consistent from the obtained pair group.

The determination unit 322 determines the quotient (similarity, correlation value, matching score) when the number of feature point pairs with the correct geometric arrangement is divided by the total number of feature points extracted from the two images to be matched. If it is higher than the threshold value Th, both are judged to be the same object.

When the determination unit 322 determines that the feature amount data R that matches the feature amount data T exists in the storage unit 33, the erasing unit 34 erases the feature amount data R that matches the feature amount data T from the storage unit 33. .. The erasing unit 34 erases the feature amount data R to be erased based on the receipt of the notification of the erasing target from the determination unit 322.

The adjusting unit 323 adjusts the value of the threshold value Th. The threshold value Th is stored in a predetermined storage area such as the storage unit 33. A specific example of the method for adjusting the threshold value Th will be described later.

The feature data T is temporarily stored in the memory at least while the authentication is performed. The feature amount data T is deleted when the authentication is completed.

FIG. 4 is a configuration diagram showing the hardware included in the image processing device 30. With reference to FIG. 4, the image processing apparatus 30 includes an arithmetic processing circuit 350, a memory 360, and a communication interface 370. The arithmetic processing circuit 350 includes a main processor 351 and an image processing dedicated processor 352.

The memory 360 includes, for example, a ROM 361, a RAM 362, and a flash memory 363. The flash memory 363 stores an operating system, an application program, and various types of data. The memory 360 (specifically, the flash memory 363) corresponds to the storage unit 33 shown in FIG. The memory 360 may be configured to include an HDD (Hard Disk Drive).

The generation unit 321 and the determination unit 322 and the erasing unit 34 shown in FIG. 3 are realized by the arithmetic processing circuit 350 executing the operating system and the application program stored in the memory 360. When executing the application program, various data stored in the memory 360 (for example, image data sent from the cameras 10 and 20) are referred to.

The main processor 351 controls the overall operation of the image processing device 30. The image processing dedicated processor 352 executes predetermined processing on the image data sent from the cameras 10 and 20. As an example, the image processing dedicated processor 352 generates the above-mentioned feature amount data from the image data. The main processor 351 performs authentication (verification) based on the feature amount data. An ASIC (Application Specific Integrated Circuit) for performing image processing may be provided instead of the image processing dedicated processor 352.

The communication interface 370 is for communicating with an external device (for example, cameras 10, 20, PLC 40). The communication interface 370 corresponds to the acquisition unit 31 of FIG.

The hardware configuration shown in FIG. 4 is an example and is not limited to this.

<C. Processing in the authentication system>
As the processing performed by the authentication system 1, the advance preparation performed before the start of the actual operation and the data amount adjustment processing which is the processing at the time of the actual operation will be described. In the following, a case where the work W is a substrate will be described as an example.

(C1. Advance preparation)
FIG. 5 is a view showing an object fingerprint in a partial region F1 of the side surface C1 of the work W. FIG. 6 is a diagram showing an object fingerprint in a partial region F2 of the surface C2 of the work W.

FIG. 7 shows the feature data of the pre-registered model substrate M, the feature data of the other three substrates X, Y, and Z, and the model substrate M when the authentication is performed using the object fingerprint on the side surface C1. It is a figure which showed the correlation with the feature quantity data acquired separately.

With reference to FIG. 7, the horizontal axis of the graph represents the number of trials, and the vertical axis of the graph represents the correlation value. In each trial, the correlation value between the feature amount data registered in advance and the feature amount data separately acquired for the model substrate M was 100%. On the other hand, the correlation value between the feature amount data registered in advance and the feature amount data of each of the other three substrates X, Y, Z, and M was around 60%.

FIG. 8 separately shows the feature data of the pre-registered model substrate M, the feature data of the other three substrates X, Y, and Z, and the model substrate when authentication is performed using the object fingerprint on the surface C2. It is a figure which showed the correlation with acquired feature quantity data.

With reference to FIG. 8, the horizontal axis of the graph represents the number of trials, and the vertical axis of the graph represents the correlation value. In each trial, the correlation value between the feature amount data registered in advance and the feature amount data separately acquired for the model substrate M was a value from the early 80% to the early 90%. Further, the correlation value between the feature amount data registered in advance and the feature amount data of each of the other three substrates X, Y, Z, and M was a value from the first half of 60% to the first half of 70%.

Comparing the correlation value shown in FIG. 7 with the correlation value shown in FIG. 8, the correlation value between the feature amount data registered in advance and the feature amount data separately acquired for the model substrate M is shown in FIG. The correlation value is higher. Therefore, it can be seen that the accuracy of authentication is higher when the object fingerprint on the side surface is used instead of the object fingerprint on the front surface of the substrate. The reason why the correlation value is higher when the object fingerprint on the side surface is used than when the object fingerprint on the surface surface is used is that the characteristics of the individual are more likely to appear than the surface which is not the cross section because the side surface is a cross section. Is.

The measurement site of the work W can be determined in advance by comparison using the correlation values as described above. The determination may be made by the user or may be made automatically by the image processing apparatus 30.

FIG. 9 is a flow chart for explaining the flow of the pre-processing process.
With reference to FIG. 9, in step S1, the portion of the work W is imaged by the camera 20. In step S2, the image processing device 30 generates feature amount data based on the image data of the object fingerprint of the relevant portion. In step S3, the image processing device 30 performs object fingerprint authentication. In step S4, the image processing device 30 determines whether or not the correlation value obtained at the time of authentication is higher than the threshold value Th.

When the image processing apparatus 30 determines that the correlation value is higher than the threshold value Th (YES in step S4), the imaging portion of the work W is appropriate for the display 60 or the like connected to the image processing apparatus 30 in step S5. Have them display that there is. When the image processing apparatus 30 determines that the correlation value is equal to or less than the threshold value Th (NO in step S4), the image processing apparatus 30 causes the image processing apparatus 30 to display in step S6 that the imaging site of the work W is not appropriate.

After step S6, the user changes the positions Pa and Pb of the cameras 10 and 20, and the feature amount data of the work W is pre-registered by the camera 10 again, and then the processes after step S1 are repeated. In the case of the above-mentioned substrate, it is preferable to set the threshold value Th to at least 80%.

(C2. Data amount adjustment processing)
Next, as described with reference to FIG. 1, a process of erasing the feature amount data R from the memory 360 (storage unit 33 in FIG. 3 and flash memory 363 in FIG. 4) will be described.

FIG. 10 is a diagram showing a flow of processing until the feature amount data R is registered in the memory 360.

With reference to FIG. 10, in step S11, the work W is imaged by the camera 10 on the upstream side. In step S12, the image processing device 30 (specifically, the generation unit 321) generates the feature amount data R of the object fingerprint based on the image data acquired from the camera 10.

In step S13, the image processing device 30 stores the generated feature amount data in the memory 360. When the next work W is conveyed to the position P1 (see FIG. 2) in step S14, the process returns to step S1 and the work W is imaged by the camera 10 on the upstream side.

FIG. 11 is a diagram showing a flow of processing from the memory 360 to erasing the feature amount data R.

When the work W is conveyed to the position P2 (see FIG. 2) with reference to FIG. 11 (YES in step S21), the camera 20 takes an image of the work W in step S22. In step S23, the image processing device 30 (specifically, the generation unit 321) generates the feature amount data T of the object fingerprint based on the image data acquired from the camera 20.

In step S24, the image processing device 30 authenticates the generated feature amount data T using the plurality of feature amount data R registered in advance. If the authentication is successful (YES in step S25), the image processing device 30 (specifically, the erasing unit 34) erases the feature data R that matches the feature data T from the memory 360. If the authentication fails (NO in step S25), the image processing apparatus 30 advances the process to step S21.

FIG. 12 is a diagram showing changes in the number of feature amount data stored in the memory 360 between the case where the data amount is adjusted by the above-mentioned erasing process and the case where the data amount is not adjusted.

With reference to FIG. 12, the horizontal axis of the graph is the number of times of imaging by the camera 10, and the vertical axis of the graph is the number of feature data R stored in the memory 360. As described above, when the erasing unit 34 executes the process of erasing the feature amount data R from the memory 360 when the authentication is successful, even if the number of work W imaged by the camera 10 increases, the memory 360 It is possible to suppress an increase in the number of feature data R accumulated in the camera.

<D. Modification example>
(D1. Authentication accuracy control processing)
A configuration example for controlling the authentication accuracy will be described.

(1) A configuration for improving the authentication accuracy by capturing the same work W a plurality of times with the repetitive identification camera 20 will be described.

FIG. 13 is a diagram showing a flow of processing from the memory 360 to erasing the feature amount data R.

When the work W is conveyed to the position P2 (see FIG. 2) with reference to FIG. 13 (YES in step S31), the camera 20 images the work W a plurality of times in step S32. In step S33, the image processing device 30 (specifically, the generation unit 321) generates a plurality of feature amount data T based on each of the plurality of image data acquired from the camera 20.

In step S34, the image processing device 30 performs an authentication process. Specifically, each of the generated plurality of feature data T is authenticated using the plurality of pre-registered feature data R. Specifically, the determination unit 322 of the image processing device 30 has generated a plurality of feature amount data T, and based on the generation of the plurality of feature amount data T, the feature amount data T of a predetermined number or more among the plurality of feature amount data T. It is determined whether or not the feature amount data R (one feature amount data R) that matches the above exists in the memory 360. The authentication processing unit 32 determines that the authentication is successful when the feature amount data R that matches the feature amount data T of the predetermined number or more exists. The predetermined number may be one or the plurality (the same value as the number of imagings).

If the authentication is successful (YES in step S35), in step S36, the image processing device 30 (specifically, the erasing unit 34) matches the feature amount data R (1) that matches the number or more of the feature amount data T determined in advance. One feature amount data R) is erased from the memory 360. If the authentication fails (NO in step S35), the image processing apparatus 30 advances the process to step S31.

According to such a configuration, the image processing device 30 performs authentication using a plurality of feature amount data, so that the authentication accuracy can be improved.

(2) Warning display regarding threshold The image processing device 30 has a case where a plurality of feature data R matching the feature data T exist in the memory 360, or a feature data R matching the feature data T exists in the memory 360. If not, the display 60 is made to perform a predetermined display.

When a plurality of feature amount data R matching the feature amount data T exist in the memory 360, the image processing device 30 causes the display 60 to display, for example, a proposal to increase the threshold value Th. Further, when the feature amount data R that matches the feature amount data T does not exist in the memory 360, the image processing device 30 causes the display 60 to display, for example, a proposal to lower the threshold value Th.

According to such a configuration, the user can set the threshold value Th to an appropriate value.

(3) Automatic Setting of Threshold Th The function of automatically setting the threshold Th to an appropriate value by the adjusting unit 323 (FIG. 3) will be described.

FIG. 14 is a flow chart showing the flow of the threshold Th setting setting process. When the work W is conveyed to the position P2 (see FIG. 2) with reference to FIG. 14 (YES in step S41), the camera 20 images the work W a plurality of times in step S42. In step S43, the image processing device 30 (specifically, the generation unit 321) generates a plurality of feature amount data T based on each of the plurality of image data acquired from the camera 20.

In step S44, the image processing device 30 performs an authentication process. Specifically, the image processing device 30 authenticates each of the generated plurality of feature amount data T using the plurality of feature amount data R registered in advance. Specifically, the determination unit 322 of the image processing device 30 counts the number of feature amount data R that matches the plurality of feature amount data T based on the generation of the plurality of feature amount data T.

When the number of matching feature amount data R is 0 (YES in step S45), the adjustment unit 323 of the image processing device 30 lowers the threshold value Th of the correlation value (similarity) by a predetermined value in step S46. After that, the image processing device 30 repeats the authentication in step S44.

When the number of matching feature amount data R is a plurality (YES in step S47), the adjustment unit 323 of the image processing device 30 raises the threshold value Th of the correlation value (similarity) by a predetermined value in step S48. After that, the image processing device 30 repeats the authentication in step S44.

When the number of matching feature data R is one (NO in step S47), the image processing device 30 stores the feature data R matching the feature data T in the memory 360 in step S49. Erase from.

As described above, when there are a plurality of feature amount data R matching the feature amount data T in the memory 360, the adjusting unit 323 arranges that the feature amount data R matching the feature amount data T becomes one. Raise the threshold Th. According to such a configuration, the threshold value Th can be set to an appropriate value.

Further, when the feature amount data R matching the feature amount data T does not exist in the memory 360, the adjustment unit 323 sets a threshold value until it is determined that one feature amount data R matching the feature amount data T exists. Lower Th. According to such a configuration, the threshold value Th can be set to an appropriate value.

Typically, when the feature amount data R that matches the feature amount data T does not exist in the memory 360, the adjusting unit 323 lowers the threshold value Th up to a predetermined lower limit value. According to such a configuration, it is possible to prevent the value of the threshold value Th from becoming equal to or less than the lower limit value.

Further, even if the threshold value reaches the lower limit value, or if the upper limit value of the threshold value Th is set, even if the threshold value Th reaches the upper limit value, matching with the feature amount data R cannot be performed yet. It is necessary to review the imaging environment. In such a case, the image processing device 30 notifies the user of a predetermined message on the screen display.

(D2. Erasing feature data R that does not match feature data T)
As described above, when the authentication is successful, the image processing device 30 erases the feature amount data R registered in advance in the memory 360. However, it is also assumed that the feature amount data R is not erased and remains in the memory 360.

For example, when the work W is extracted in the middle of the route, the work W is not imaged by the camera 20 on the downstream side unless the work W is returned to its original position. In such a case, since the feature amount data T that matches the feature amount data R is not generated, the feature amount data R continues to remain in the memory 360. The existence of such feature amount data R can be a factor that increases the possibility of erroneous authentication and an increase factor in the number of matchings.

Therefore, it is preferable to configure the image processing device 30 so as to erase the feature amount data R remaining in the memory 360 for a predetermined period (for example, several days) or longer. In particular, from the viewpoint of traceability, the fact that the feature amount data R remains means that a shortage has occurred in the process. Therefore, not only the feature amount data R is deleted but also the operator is notified by the screen display. It is preferable to do so. As such a notification, the image processing device 30 preferably displays the date, time, minute, and second of the imaging of the feature amount data R on the display 60 in a list display format or the like.

[Embodiment 2]
In the first embodiment, a configuration in which an object fingerprint of one part (part B1) of the work W is imaged by the camera 10 on the upstream side and the camera 20 on the downstream side has been described as an example. That is, an example is given of a configuration in which authentication is performed using an object fingerprint of one part of the work W.

In the present embodiment, a configuration in which authentication is performed by fingerprints of objects at a plurality of parts of the work W will be described. As an example, a configuration in which authentication is performed using object fingerprints of two parts will be described. More specifically, the configuration in which the authentication as a whole is successful will be described on the condition that the authentication is successful in the two parts.

FIG. 15 is a diagram showing a specific example of the authentication system 1A.
With reference to FIG. 15, the authentication system 1A includes cameras 10, 11, cameras 20, 21, an image processing device 30, a PLC 40, and a server device 50. The cameras 10 and 11 and the cameras 20 and 21 are communicably connected to the image processing device 30.

The camera 11 refers to a predetermined portion of a plurality of work Ws, each of which is conveyed along a predetermined route, and a portion different from the portion B1 (hereinafter, referred to as "site B2") of the route. The image is taken at the position Pc on the upstream side. The camera 21 images the portion B2 of these objects at the position Pd on the downstream side of the path.

In this way, the cameras 10 and 20 image the portion B1 of the work W transported on the route by the transport device 900, and the cameras 11 and 21 capture the portion B2 of the work W transported on the route by the transport device 900. To image. The cameras 10, 11, 20, and 21 transmit the image data obtained by imaging to the image processing device 30.

FIG. 16 is a diagram for explaining the positions of the portion B1 and the portion B2. With reference to FIG. 16, as an example, the portion B1 can be the first side surface of the work W, and the portion B2 can be the side surface opposite to the first side surface.

FIG. 17 is a diagram showing a flow of processing until the feature amount data R and R'related to the two parts B1 and B2 are registered in the memory 360.

In step S51, the image of the work W portion B1 is performed by the camera 10 on the upstream side. In step S52, the image processing device 30 generates the feature amount data R of the object fingerprint at the portion B1 based on the image data acquired from the camera 10. In step S53, the image processing device 30 stores the generated feature amount data in the memory 360.

In step S54, the camera 11 takes an image of the work W portion B2. In step S55, the image processing device 30 generates the feature amount data R'of the object fingerprint at the portion B2 based on the image data acquired from the camera 11. In step S56, the image processing apparatus 30 stores the generated feature amount data in the memory 360 in association with the feature amount data R generated in step S52.

When the next work W is conveyed to the position P1 (see FIG. 2) in step S57, the process returns to step S51, and the image of the work W by the camera 10 on the upstream side is repeated.

FIG. 18 is a diagram showing a processing flow from the memory 360 to erasing the feature amount data R, R'.

When the work W is conveyed to the position P2 (see FIG. 2) with reference to FIG. 18 (YES in step S61), the camera 20 takes an image of the portion B1 of the work W in step S62. In step S63, the image processing device 30 generates the feature amount data T of the object fingerprint in the portion B1 based on the image data acquired from the camera 20.

In step S64, the camera 21 takes an image of the portion B2 of the work W. In step S65, the image processing device 30 generates the feature amount data T'of the object fingerprint at the portion B2 based on the image data acquired from the camera 21.

In step S66, the image processing device 30 performs an authentication process. Specifically, the image processing device 30 authenticates the generated feature data T, T'using a plurality of pre-registered feature data R, R'. Specifically, the determination unit 322 of the image processing device 30 determines that the feature amount data R that matches the generated feature amount data T exists in the memory 360, and the feature amount data that matches the generated feature amount data T'. It is determined whether or not R'exists in the memory 360. When the feature amount data R that matches the feature amount data T exists in the memory 360 and the feature amount data R'that matches the feature amount data T'exists, the determination unit 322 determines the feature amount data R and the feature amount data. It is judged that the authentication is successful on the condition that it is associated with R'.

When the authentication is successful (YES in step S67), the image processing device 30 erases the feature amount data R matching the feature amount data T and the feature amount data R'matching the feature amount data T'from the memory 360. To do. If the authentication fails (NO in step S67), the image processing apparatus 30 advances the process to step S61.

According to such a configuration, since the image processing device 30 performs authentication using feature amount data of a plurality of parts, it is possible to improve the authentication accuracy.

In the first and second embodiments, the case where the object to be authenticated is the work W transported by the transport device 900 has been described as an example, but the present invention is not limited to this. For example, it may be a product or the like that is loaded and unloaded in a warehouse or the like.

<Additional notes>
[Structure 1] A first position (Pa) on the upstream side of a plurality of objects (W), each of which is transported along a predetermined path, is imaged at a first position (Pa) on the upstream side of the path. 1 camera (10) and
A second camera (20) that captures the first portion (B1) of each of the objects (W) at a second position (Pb) on the downstream side of the path.
An image processing device (30) communicatively connected to the first camera (10) and the second camera (20) is provided.
The image processing device (30) represents the feature amount of the object fingerprint of the first part (B1) based on the image data of the first part (B1) acquired from the first camera (10). A generation means (321) for generating the first feature amount data (R) for each object (W), and
A storage means (33,360) for storing each of the generated first feature amount data (R) is included.
The generation means (321) represents the feature amount of the object fingerprint of the first part (B1) based on the image data of the first part (B1) acquired from the second camera (20). The second feature amount data (T) is further generated for each object (W), and the second feature amount data (T) is further generated.
The image processing device (30)
Based on the generation of the second feature amount data (T), the first feature amount data (R) that matches the second feature amount data (T) is the storage means (33,360). Judgment means (322) for determining whether or not it exists in
An authentication system (1) further including an erasing means (34) for erasing the first feature amount data (R) that matches the second feature amount data (T) from the storage means (33,360). ..

[Structure 2] A third camera (11) that captures a second portion (B2) of each object (W) at a third position (Pc) on the upstream side of the path.
A fourth camera (21) that captures a second portion (B2) of each object (W) at a fourth position (Pd) downstream of the path is further provided.
The generation means (321) further
Based on the image data of the second part (B2) acquired from the third camera (11), the third feature amount data (R) representing the feature amount of the object fingerprint of the second part (B2). ') Is generated for each object (W).
Based on the image data of the second part (B2) acquired from the fourth camera (21), the fourth feature amount data (T) representing the feature amount of the object fingerprint of the second part (B2). ') Is generated for each object (W).
The storage means (33,360) stores the first feature amount data (R) and the third feature amount data (R') in association with each object (W).
The determination means (322) is based on the generation of the fourth feature amount data (T'), and the third feature amount data (R) that matches the fourth feature amount data (T'). Further determining whether') exists in the storage means (33,360),
In the erasing means (34), the first feature amount data (R) that matches the second feature amount data (T) is present in the storage means (33,360), and the fourth feature amount data (T) is present. When the third feature amount data (R') that matches the feature amount data (T') exists in the storage means (33,360), the first feature amount data (R) and the third feature amount data (R) The storage means (33,360) stores the first feature amount data (R) and the third feature amount data (R') on condition that the feature amount data (R') is associated with the feature amount data (R'). The authentication system (1A) according to configuration 1, which is erased from.

[Structure 3] The second camera (20) takes images of each object (W) a plurality of times.
Based on the plurality of imagings, the generation means (321) obtains the second feature amount data (T) representing the feature amount of the object fingerprint of the first part (B1) into the object (W). Generate multiple for each,
The determination means (322) is based on the generation of the plurality of second feature amount data (T), and is equal to or more than a predetermined number of the plurality of second feature amount data (T). It is determined whether or not the first feature amount data (R) that matches the second feature amount data (T) exists in the storage means (33,360).
The erasing means (34) erases the first feature amount data (R) that matches the predetermined number or more of the second feature amount data (T) from the storage means (33, 360). The authentication system (1) according to the configuration 1 or 2.

[Structure 4] In the determination means (322), the degree of similarity between the second feature amount data (T) and the first feature amount data (R) is equal to or higher than a predetermined threshold value (Th). Is a condition, and it is determined that the first feature amount data (R) that matches the second feature amount data (T) exists in the storage means (33,360).
The image processing apparatus (30) further includes an adjusting means (323) for adjusting the threshold value (Th).
When a plurality of the first feature amount data (R) matching the second feature amount data (T) are present in the storage means (33,360), the adjusting means (323) is said to be the first. The authentication system according to any one of configurations 1 to 3, which raises the threshold value (Th) so that the first feature amount data (R) that matches the feature amount data (T) of 2 becomes one. (1).

[Structure 5] When the first feature amount data (R) that matches the second feature amount data (T) does not exist in the storage means (33,360), the adjusting means (323) The authentication system according to configuration 4, wherein the threshold value (Th) is lowered until it is determined that one of the first feature amount data (R) that matches the second feature amount data (T) exists. 1).

[Structure 6] When the first feature amount data (R) that matches the second feature amount data (T) does not exist in the storage means (33,360), the adjusting means (323) The authentication system (1) according to the configuration 5, wherein the threshold value (Th) is lowered up to a predetermined lower limit value.

[Structure 7] A display device (60) communicatively connected to the image processing device (30) is further provided.
In the image processing apparatus (30), when a plurality of the first feature amount data (R) matching the second feature amount data (T) are present in the storage means (33,360), or the first feature amount data (T). When the first feature amount data (R) that matches the feature amount data (T) of 2 does not exist in the storage means (33, 360), a predetermined display is displayed on the display device (60). The authentication system (1) according to any one of configurations 1 to 3 to be performed.

[Structure 8] A predetermined portion (B1) of a plurality of objects (W), each of which is transported along a predetermined path, is imaged at a first position (Pa) on the upstream side of the path. A second camera (20) that captures the first camera (10) and the predetermined portion (B1) of each object (W) at a second position (Pb) on the downstream side of the path. A data processing method in an authentication system (1) including the first camera (10) and an image processing device (30) communicatively connected to the second camera (20).
First feature amount data representing the feature amount of the object fingerprint of the predetermined part (B1) based on the image data of the predetermined part (B1) acquired from the first camera (10). A step of generating (R) for each object (W), and
A step of storing each of the generated first feature amount data (R) in the memory (360, 33), and
Second feature amount data representing the feature amount of the object fingerprint of the predetermined part (B1) based on the image data of the predetermined part (B1) acquired from the second camera (20). A step of generating (T) for each object (W), and
Based on the generation of the second feature amount data (T), the first feature amount data (R) that matches the second feature amount data (T) is stored in the memory (360, 33). Steps to determine if it exists and
A data processing method comprising a step of erasing the first feature amount data (R) that matches the second feature amount data (T) from the memory (360, 33).

It should be considered that the embodiments disclosed this time are exemplary in all respects and not restrictive. The scope of the present invention is shown by the scope of claims rather than the description of the embodiment described above, and is intended to include all modifications within the meaning and scope equivalent to the scope of claims.

1,1A authentication system, 10,11,20,21 camera, 30 image processing unit, 31 acquisition unit, 32 authentication processing unit, 33 storage unit, 34 erasing unit, 50 server device, 60 display, 321 generator, 322 judgment Unit, 323 adjustment unit, 350 arithmetic processing circuit, 351 main processor, 352 image processing dedicated processor, 360 memory, 361 ROM, 362 RAM, 363 flash memory, 370 communication interface, 900 carrier, B1, B2 part, C1 side surface, C2 surface, P1, P2, Pa, Pb, Pc, Pd position, W work.

Claims (8)

A first camera that captures a first portion of a plurality of objects, each of which is transported along a predetermined path, at a first position on the upstream side of the path.
A second camera that captures the first portion of each object at a second position downstream of the path.
An image processing device communicatively connected to the first camera and the second camera is provided.
The image processing device is
A generation means for generating first feature amount data representing the feature amount of the object fingerprint of the first part for each object based on the image data of the first part acquired from the first camera. ,
Including a storage means for storing each of the generated first feature amount data.
The generation means generates second feature amount data representing the feature amount of the object fingerprint of the first part for each object based on the image data of the first part acquired from the second camera. Generate more and
The image processing device is
Based on the generation of the second feature amount data, a determination means for determining whether or not the first feature amount data matching the second feature amount data exists in the storage means, and
An authentication system further comprising an erasing means for erasing the first feature amount data that matches the second feature amount data from the storage means. A third camera that captures a second portion of each object at a third position upstream of the path.
A fourth camera that captures a second portion of each object at a fourth position downstream of the path is further provided.
The generation means further
Based on the image data of the second part acquired from the third camera, a third feature amount data representing the feature amount of the object fingerprint of the second part is generated for each object.
Based on the image data of the second part acquired from the fourth camera, a fourth feature amount data representing the feature amount of the object fingerprint of the second part is generated for each object.
The storage means stores the first feature amount data and the third feature amount data in association with each other for each object.
Based on the generation of the fourth feature amount data, the determination means further determines whether or not the third feature amount data matching the fourth feature amount data exists in the storage means.
In the erasing means, the first feature amount data that matches the second feature amount data exists in the storage means, and the third feature amount data that matches the fourth feature amount data is present. When present in the storage means, the first feature amount data and the third feature amount data are provided on the condition that the first feature amount data and the third feature amount data are associated with each other. The authentication system according to claim 1, wherein the data is erased from the storage means. The second camera takes images of each object a plurality of times.
The generation means generates a plurality of second feature amount data representing the feature amount of the object fingerprint of the first part for each object based on the plurality of imagings.
The determination means coincides with the second feature amount data of a predetermined number or more among the plurality of second feature amount data based on the generation of the plurality of second feature amount data. It is determined whether or not the first feature amount data to be stored exists in the storage means.
The authentication system according to claim 1 or 2, wherein the erasing means erases the first feature amount data that matches the predetermined number or more of the second feature amount data from the storage means. The determination means matches the second feature amount data on the condition that the similarity between the second feature amount data and the first feature amount data is equal to or more than a predetermined threshold value. Judging that the feature amount data of 1 exists in the storage means,
The image processing apparatus further includes an adjusting means for adjusting the threshold value.
When a plurality of the first feature amount data matching the second feature amount data are present in the storage means, the adjusting means matches the first feature amount data with the second feature amount data. The authentication system according to any one of claims 1 to 3, which raises the threshold value so that the data becomes one. The adjusting means has the first feature amount data that matches the second feature amount data when the first feature amount data that matches the second feature amount data does not exist in the storage means. The authentication system according to claim 4, wherein the threshold value is lowered until it is determined that one is present. The adjusting means lowers the threshold value up to a predetermined lower limit value when the first feature amount data matching the second feature amount data does not exist in the storage means. The authentication system according to 5. A display device communicatively connected to the image processing device is further provided.
In the image processing device, when a plurality of the first feature amount data matching the second feature amount data are present in the storage means, or the first feature amount matching the second feature amount data. The authentication system according to any one of claims 1 to 3, which causes the display device to perform a predetermined display when the data does not exist in the storage means. A first camera that captures a predetermined portion of a plurality of objects, each of which is transported along a predetermined path, at a first position on the upstream side of the path, and the predetermined portion of each object. Authentication including a second camera that images the formed portion at a second position on the downstream side of the path, and an image processing device communicably connected to the first camera and the second camera. It is a data processing method in the system
A step of generating first feature amount data representing the feature amount of the object fingerprint of the predetermined part for each object based on the image data of the predetermined part acquired from the first camera. When,
A step of storing each of the generated first feature amount data in the memory, and
A step of generating second feature amount data representing the feature amount of the object fingerprint of the predetermined part for each object based on the image data of the predetermined part acquired from the second camera. When,
Based on the generation of the second feature amount data, a step of determining whether or not the first feature amount data matching the second feature amount data exists in the memory, and
A data processing method comprising a step of erasing the first feature amount data that matches the second feature amount data from the memory.

Images