KR20240053954A - Apparatus for Environment Adaptive Machine Vision and Driving Method Thereof - Google Patents

Abstract

The present invention relates to an environmentally adaptive machine vision device and a method of driving the device. The environmentally adaptive machine vision device according to an embodiment of the present invention includes a communication interface unit for receiving photographed data of an inspected product, and a received photographic image. A control unit that analyzes data and detects changes in the rate of non-good condition (NG) of the product based on the analysis results, judges changes in the shooting environment of the inspected product based on the detected results, and corrects the analysis results based on the judgment results. It can be included.

Description

Environmental adaptive machine vision device and driving method of the device {Apparatus for Environment Adaptive Machine Vision and Driving Method Thereof}

The present invention relates to an environmentally adaptive machine vision device and a method of operating the device. More specifically, in the field of machine vision, which is widely used for product inspection in production lines such as factories, Environmentally adaptive machine vision devices and devices that detect (environmental) changes by recognizing changes in the distribution of good (NG) and implement machine vision technology adaptively (or flexibly) based on the detection results of environmental changes during product inspection. It is about the driving method of .

Recently, industrial automation is gradually accelerating as elements such as machine vision, robots, sensors, PLC, and computers are combined across industrial fields such as automobiles, electronic devices, displays, semiconductors, food and beverage, and sports. there is. As industrial automation accelerates across industrial sectors, various technologies are being introduced to automate the detailed inspection process that is performed on mass-produced products before they are shipped. Vision inspection (machine vision inspection), one of these detailed inspections, replaces the existing method of human judgment with the naked eye, and uses images taken by a camera to check, for example, the labeling status of a product, the printing status of packaging, or the printing status of barcodes. Check your back.

Specifically, vision inspection is performed by photographing the surface of the inspection object to obtain a surface image and comparing the obtained surface image with a reference image to check for defects in the inspection object. Devices for vision inspection are installed throughout the process line and perform vision inspection on products (hereinafter referred to as inspection objects) moving along the process line. In other words, manufacturers are forming large-scale production lines, and installing multiple vision equipment, such as IP cameras, on each production line to inspect defect rates and quality for each process.

However, conventional vision inspection has a problem in that it cannot guarantee the accuracy of the existing model, that is, the inspection program, when the product inspection environment, such as the intensity of lighting or location, changes. Therefore, field workers must take arbitrary action, but since there are no clear standards, it is difficult to guarantee the accuracy of existing models.

Additionally, conventionally, when the vision inspection environment changes, a new model must be learned to suit the changed environment, but this has the problem of being a burden in terms of cost.

Korean Patent Publication No. 10-2022496 (2019.09.10) Korea Patent Publication No. 10-2022-0085115 (2022.06.22) Korean Patent Publication No. 10-2022-0085589 (2022.06.22)

An embodiment of the present invention detects (environmental) changes by recognizing changes in the distribution of good (OK) and bad (NG) during the product inspection process in the field of machine vision, which is widely used for product inspection in production lines such as factories. The purpose is to provide an environmentally adaptive machine vision device and a method of operating the device to implement machine vision technology adaptively (or flexibly) based on the detection results of environmental changes during product inspection.

An environmentally adaptive machine vision device according to an embodiment of the present invention includes a communication interface unit that receives photographed data of an inspected product, and a communication interface unit that analyzes the received photographed data and determines the non-good state (NG) of the product based on the analysis results. It includes a control unit that detects a change in rate, determines a change in the photographing environment of the inspected product according to the detected result, and corrects the analysis result based on the judgment result.

The control unit may determine the unhealthy state based on a preset threshold and correct the analysis result by automatically adjusting the preset threshold based on the determination result.

The control unit may automatically set the threshold value higher than the preset threshold value when the ratio of the unhealthy state increases, and automatically set the threshold value lower than the preset threshold value when the ratio of the unhealthy state decreases. .

The control unit may automatically adjust a preset threshold by determining environmental changes by taking into account changes in the intensity or location of lighting in a space where the product is inspected or changes in the location of the product as changes in the photographing environment.

The control unit may calculate the analysis result of the received photographed data into a score and compare the calculated score with a preset threshold to determine the unhealthy state.

In addition, a method of driving an environment-adaptive machine vision device according to an embodiment of the present invention includes a communication interface unit receiving photographed data of an inspected product, and a control unit analyzing the received photographed data and based on the analysis result. Detecting a change in the ratio of the non-good state (NG) of the product, determining a change in the photographing environment of the inspected product according to the detection result, and correcting the analysis result based on the judgment result.

In the correction step, the unhealthy state is determined based on a preset threshold, and the analysis result can be corrected by automatically adjusting the preset threshold based on the determination result.

In the step of correcting, when the ratio of the unhealthy state increases, the threshold value is automatically set higher than the preset threshold value, and when the ratio of the unhealthy state decreases, the threshold value is automatically set lower than the preset threshold value. You can.

In the correction step, a preset threshold value can be automatically adjusted by determining environmental changes in consideration of changes in the intensity or location of lighting in the space where the product is inspected or changes in the location of the product, which are changes in the photographing environment.

In the correction step, the analysis result of the received photographed data may be calculated into a score, and the calculated score may be compared with a preset threshold to determine the unhealthy state.

According to an embodiment of the present invention, the accuracy of product inspection can be guaranteed even if the product inspection environment, such as the lighting intensity or location of the space where product inspection is performed at the machine vision site or the location of the product, varies.

In addition, embodiments of the present invention can be used in various environments without new learning, such as a new model, such as a deep learning program, in accordance with the changed environment of product inspection.

1 is a diagram showing an environment-adaptive machine vision system according to an embodiment of the present invention;
Figure 2 is a block diagram illustrating the detailed structure of the environmentally adaptive machine vision device of Figure 1;
Figures 3 to 5 are flowcharts of the environment adaptive machine vision operation process, and
FIG. 6 is a flowchart showing the operation process of the environmentally adaptive machine vision device of FIG. 1.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

Figure 1 is a diagram showing an environment-adaptive machine vision system according to an embodiment of the present invention.

As shown in FIG. 1, the environmentally adaptive machine vision system 90 according to an embodiment of the present invention includes part or all of the imaging device 100, the communication network 110, and the environmentally adaptive machine vision device 120. Includes.

Here, “including part or all” means that some components such as the communication network 110 are omitted and the imaging device 100 and the environmentally adaptive machine vision device 120 perform direct communication (e.g., P2P, etc.). This means that some or all of the components that make up the environmentally adaptive machine vision device 120 can be integrated into a network device (e.g., wireless exchange device, etc.) that makes up the communication network 110. , it is explained as including everything in order to facilitate a sufficient understanding of the invention.

The imaging device 100 may include various types of cameras installed for product inspection at a production line, such as a factory, to perform machine vision operations. Here, the imaging device 100 can increase accuracy when inspecting products by using a high-resolution CCD camera. Here, machine vision is a method for image-based automatic inspection and analysis. The imaging device 100 according to an embodiment of the present invention is located on a conveyor belt of a production line where product inspection is performed, and is mounted on a camera installed on the upper side of a random position when the product to be inspected is loaded on the conveyor belt and moves. You can perform shooting operations by

In this process, the photographing device 100 may provide the photographed image of the product to the environmentally adaptive machine vision device 120 via the communication network 110 of FIG. 1. Of course, the imaging device 100 may also be linked to an edge device (not shown) installed around it or, for example, around a conveyor belt. The edge device analyzes the captured image of the photographing device 100, omits the analysis of pixel data corresponding to the background, extracts and provides only the portion corresponding to the product, or provides the analysis result to the environmentally adaptive machine vision device 120. can be provided. In this case, the data processing load of the communication network 110 is reduced, so the data processing speed in the environmentally adaptive machine vision device 120 can rapidly increase.

The communication network 110 includes both wired and wireless communication networks. For example, a wired or wireless Internet network may be used or linked as the communication network 110. Here, the wired network includes Internet networks such as cable networks and public switched telephone networks (PSTN), and the wireless communication network includes CDMA, WCDMA, GSM, EPC (Evolved Packet Core), LTE (Long Term Evolution), and Wibro networks. It means including. Of course, the communication network 110 according to an embodiment of the present invention is not limited to this, and can be used as an access network for a next-generation mobile communication system to be implemented in the future, for example, a cloud computing network in a cloud computing environment, a 5G network, etc. For example, if the communication network 110 is a wired communication network, the access point within the communication network can connect to the telephone company's exchange office, etc., but in the case of a wireless communication network, data is processed by connecting to the SGSN or GGSN (Gateway GPRS Support Node) operated by the communication company, or Data can be processed by connecting to various repeaters such as BTS (Base Transceiver Station), NodeB, and e-NodeB.

Communication network 110 may also include access points. Access points include small base stations such as femto or pico base stations that are often installed in buildings. Here, femto or pico base stations are classified according to the maximum number of imaging devices 100, etc., that can be connected in the classification of small base stations. Of course, the access point may include a short-distance communication module for performing short-distance communication such as Zigbee and Wi-Fi with the photographing device 100, etc. Access points can use TCP/IP or RTSP (Real-Time Streaming Protocol) for wireless communication. Here, in addition to Wi-Fi, short-range communication can be performed using various standards such as Bluetooth, Zigbee, infrared (IrDA), RF (Radio Frequency) such as UHF (Ultra High Frequency) and VHF (Very High Frequency), and ultra-wideband communication (UWB). You can. Accordingly, the access point extracts the location of the data packet, specifies the best communication path for the extracted location, and forwards the data packet along the designated communication path to the next device, such as the adaptive machine vision device 120. . Access points can share multiple lines in a typical network environment and include, for example, routers, repeaters, and repeaters.

The environmentally adaptive machine vision device 120 changes the distribution of good (OK) or bad (NG) during the product inspection process based on the product captured image provided by the photographing device 100 of FIG. 1 or the analysis result of the photographed image. , detects environmental changes, etc. Here, the environment refers to the environment where the imaging device 100 is installed in the production line and product inspection takes place, and environmental changes refer to changes in the strength or location of the product in the space and the location of the product on the conveyor belt, etc. can do. For example, the environmentally adaptive machine vision device 120 detects relatively sensitive changes in the NG (No Good) ratio. Additionally, the environmentally adaptive machine vision device 120 automatically adjusts the threshold value according to the changed environment. For example, if the NG ratio increases, the threshold can be raised, and if the NG ratio decreases, the threshold can be lowered. In other words, the size of the threshold can be increased or decreased based on the preset threshold. In this way, the environmentally adaptive machine vision device 120 re-determines the results from the results of the part where the environmental change is detected to a new threshold value, that is, the threshold value of the changed state by adapting to the environment. Accordingly, it can be used in various environments without new learning such as deep learning programs.

As will be discussed in detail later, the environmentally adaptive machine vision device 120 can receive a captured image provided from the photographing device 100. This captured image may be one image, but may be provided with a plurality of captured images. Of course, it could be a video. The environmentally adaptive machine vision device 120 can extract and analyze images of product parts in the process of analyzing received captured images. In addition, comparison results can be generated by comparing the product's standard image and, more precisely, the pixel data of the standard image. Here, the comparison result can be calculated as a score. For example, when calculating the score, the total score is set to 10 or 100, the score for the product is calculated based on the comparison results of the pixel data, and the score is compared with the standard threshold to determine whether it is good or bad.

And the environmentally adaptive machine vision device 120 also calculates the ratio of good to bad. Specifically, the environmentally adaptive machine vision device 120 can set a standard value for the ratio of good and bad, and therefore determines whether the bad rate exceeds the standard value or whether the environment of the space where product inspection is performed has changed. For example, the environmentally adaptive machine vision device 120 can determine the characteristics of lighting and determine the location of a product. For example, in the case of lighting, the background of the input captured image can be determined, and for this, the color or gradation value of the background can be determined to determine whether the current lighting environment has changed. Of course, this is related to the intensity of lighting, etc. In addition, it can be determined whether the installation location of the lighting has changed.

Simply put, if the lighting (device) implements IoT technology and enables short-distance communication such as Bluetooth, the photographing device 100 attempts to communicate with the lighting device at the time of determining a change in the environment and receives values such as location information. The adaptive machine vision device 120 can use the location information of the lighting device provided by the imaging device 100 or the lighting device. In addition, the environmentally adaptive machine vision device 120 can determine the positions of products subject to inspection, for example, located on a conveyor belt. For example, by analyzing captured images, it may be determined that there is an abnormality in the operating state of the conveyor belt when it is determined that the position of the product has changed beyond a certain time or standard number of times. Therefore, the environmentally adaptive machine vision device 120 can take measures to ensure accurate product inspection by reflecting the abnormal operating state unless the conveyor belt is broken. As mentioned earlier, the percentage of unhealthy states increases, so the threshold value is raised.

Of course, in the actual environment of the production line, the domain, or product location area, changes in various ways during the production process depending on the product, so it may be difficult to quantitatively define the relevant part. In addition, even in the same product, the domain may change due to various factors that are difficult to predict. For example, in the semiconductor process, the aspect of the captured image may change not only depending on environmental factors (e.g., lighting or replacement of consumables, etc.) but also simply on the condition of the wafer. there is. Therefore, because it is difficult to respond to all aspects of these domain changes, the environmentally adaptive machine vision device 120 according to an embodiment of the present invention simply uses a method of inferring the threshold using inference or statistics of the inference results. It can be done. Additionally, since changes can occur frequently, changes can be detected and adjusted by automatically calculating statistics.

In order to perform the above operation, the environmentally adaptive machine vision device 120 according to an embodiment of the present invention can first calculate the total NG ratio and also calculate the current NG ratio. And whether the domain has changed can be determined through the overall NG ratio and the current NG ratio. That is, when there is a change in the environment, such as a domain, the environmentally adaptive machine vision device 120 automatically adjusts the preset threshold value, and when the NG ratio increases, it increases the preset threshold value according to the environmental change. In other words, it can be set larger than the standard threshold. Additionally, if the NG ratio decreases, the preset threshold is lowered according to environmental changes.

As a result of the above configuration, the embodiment of the present invention can increase the accuracy of product inspection by adaptively responding to the changing environment even when the environment changes in a random space, such as a production line where product inspection (or inspection) is performed. In addition, existing machine vision programs, such as artificial intelligence deep learning programs, may be able to be used continuously in various environments through simple development of the program without new learning.

Figure 2 is a block diagram illustrating the detailed structure of the environmentally adaptive machine vision device of Figure 1;

As shown in Figure 2, the environmentally adaptive machine vision device 120 according to an embodiment of the present invention includes a communication interface unit 200, a control unit 210, an environmentally adaptive machine vision unit 220, and a storage unit ( 230) includes part or all of

Here, “including part or all” means that the environmentally adaptive machine vision device 120 is configured by omitting some components such as the storage unit 230 or some components such as the environmentally adaptive machine vision unit 220. This means that the element can be integrated into other components such as the control unit 210, and is described as being fully included to facilitate a sufficient understanding of the invention.

The communication interface unit 200 communicates with each of the plurality of photographing devices 100 via the communication network 110 of FIG. 1. Here, the plurality of photographing devices 100 may include cameras installed in various places such as a production line to inspect products. The communication interface unit 200 may perform operations such as modulation/demodulation, encoding/decoding, and muxing/demuxing during communication. Since this is obvious to those skilled in the art, further explanation is omitted.

For example, the communication interface unit 200 includes a device for analyzing images in the photographing device 100, or when linking a separate image analysis device such as an edge device, the photographing device 100 is connected to the photographing device 100 from the device. Captured images can be received. Of course, a vision inspection can be a single still image, but it can also be a video image captured in real time. The communication interface unit 200 receives the images or analysis results of the images and provides them to the control unit 210.

The control unit 210 performs the overall control operation of the communication interface unit 200, the environmentally adaptive machine vision unit 220, and the storage unit 230 of FIG. 2, which constitute the environmentally adaptive machine vision device 120 of FIG. 1. In charge. Typically, the control unit 210 receives a captured image related to product inspection provided by the communication interface unit 200 or an analysis result of the image, stores it temporarily in the storage unit 230, and then retrieves it from the environmentally adaptive machine vision unit ( 220).

Additionally, the control unit 210 controls the communication interface unit 200 to store data processed in a specified format in the DB 120a of FIG. 1 according to a request from the environmentally adaptive machine vision unit 220. In addition, at the request of the environmentally adaptive machine vision unit 220, the control unit 210 further receives various data related to environmental changes in the space where product inspection is performed from the photographing device 100 or lighting device of FIG. 1, and stores this data in the environment. It is provided to the adaptive machine vision unit 220. For example, when a request is made directly from the environmentally adaptive machine vision device 120 to a lighting device, the lighting device may provide its current location information or posture information.

The environmentally adaptive machine vision unit 220 performs product inspection operations by applying machine vision technology to ensure that product inspection is performed faster and more accurately than with the naked eye in production lines, etc. The environmentally adaptive machine vision unit 220 can analyze the image of the product captured by the photographing device 100 of FIG. 1 and calculate a score based on the analysis results. Then, the calculated score can be compared with the standard value to determine whether the product is good or bad. In other words, if the calculated value exceeds the standard value, it can be good, and if the calculated value does not exceed the standard value, it can be bad.

The environment of the space where product inspection takes place may change intermittently. Typically, the lighting characteristics of the inspection space can change, and when product inspection is performed while moving the product on a conveyor belt, the operating status of the conveyor belt can also affect environmental changes. Therefore, the environmentally adaptive machine vision unit 220 not only determines whether it is good or bad, but also checks the ratio, or distribution, between good and bad. Therefore, if the ratio of unsatisfactory conditions increases at some point, it is determined whether the environment has changed. For example, the state of lighting can be determined by analyzing the pixel data of the background image through the received captured image. Alternatively, the photographing posture of the photographing device 100 may be determined by receiving data related to the posture of the photographing device 100. Furthermore, various data related to the location information or operating status of the lighting device can be received from the lighting device or via the photographing device 100, and the environment of the product inspection space can be determined based on this.

Of course, in the embodiment of the present invention, rather than periodically monitoring the environment of the space, the product is classified as good or bad based on a preset threshold, and the proportion, that is, the distribution ratio, is monitored at the same time. When non-good results begin to increase, environmental changes in the product inspection space are judged based on this. Of course, in this process, it may be possible to predict the timing of environmental changes more accurately by applying programs such as artificial intelligence. For example, if the position of the product is constantly changing due to the aging of the conveyor belt, aging can be predicted by learning the operation data of the conveyor device. This may change the position of the inspected product, so take appropriate action in advance through prediction. can be taken.

The environmentally adaptive machine vision unit 220 according to an embodiment of the present invention determines whether the environment has changed at some point, that is, when the ratio of non-good products (suddenly) increases as a result of product inspection, and determines that the environment has changed as a result of the determination. When this happens, adjust the preset judgment criteria thresholds for good and bad. In other words, when the standard for dividing good and bad is considered to be a total of 10, if the standard value was set at 5, by raising the standard value to 6, products that are in the bad category can be brought into the good category. . Of course, the automatic adjustment of the preset reference value, that is, the threshold, can be automatically adjusted differently for each variable through experiment or by collecting and analyzing data. For example, if the standard value is automatically adjusted to 6 due to a belt aging problem, lighting may be automatically adjusted to 7, etc.

The storage unit 230 temporarily stores various types of data processed under the control of the control unit 210. For example, the storage unit 230 can store pixel data for images captured by the photographing device 100, but can also temporarily store device identification information of the photographing device 100. Accordingly, it can be determined with which photographing device 100 a specific product was photographed, and therefore, if it is determined that there is a problem with the photographing device 100, it may be replaced. In this regard, device identification information can be used.

In addition to the above, the communication interface unit 200, control unit 210, environmentally adaptive machine vision unit 220, and storage unit 230 of FIG. 2 can perform various operations, and other details are fully described above. So, I would like to replace it with those contents.

The communication interface unit 200, control unit 210, environmentally adaptive machine vision unit 220, and storage unit 230 of FIG. 2 according to an embodiment of the present invention are composed of hardware modules that are physically separated from each other, but each The module may store software to perform the above operations internally and execute it. However, the software is a set of software modules, and each module can be formed as hardware, so there will be no particular limitation on the configuration of software or hardware. For example, the storage unit 230 may be hardware, such as storage or memory. However, since it is possible to store information through software (repository), the above content will not be specifically limited.

Meanwhile, as another embodiment of the present invention, the control unit 210 may include a CPU and memory, and may be formed as a single chip. The CPU includes a control circuit, an arithmetic unit (ALU), an instruction interpretation unit, and a registry, and the memory may include RAM. The control circuit performs control operations, the operation unit performs operations on binary bit information, and the command interpretation unit includes an interpreter or compiler, which can convert high-level language into machine language and machine language into high-level language. , the registry may be involved in software data storage. According to the above configuration, for example, at the beginning of operation of the environmentally adaptive machine vision device 120, the program stored in the environmentally adaptive machine vision unit 220 is copied and loaded into memory, that is, RAM, and then executed. The data operation processing speed can be rapidly increased. In the case of deep learning models, they can be loaded into GPU memory rather than RAM and executed by accelerating the execution speed using GPU.

3 to 5 are flowcharts of an environment-adaptive machine vision operation process.

For convenience of explanation, referring to FIGS. 3 to 5 together with FIG. 1, the environmentally adaptive machine vision device 120 of FIG. 1 according to an embodiment of the present invention uses a captured image of the inspected product provided by the photographing device 100 or Receive video (S300).

Next, the environmentally adaptive machine vision device 120 analyzes the received captured image or video and calculates a score according to a preset method (S310). For example, you can extract only the domain (region) corresponding to the product from the captured image and then analyze the pixel data for the product in the extracted area. Since the standard pixel data for the inspected product is pre-stored, the number of pixels that do not match each other can be calculated by comparing the pixel values of the pixels with the standard value, and the score can be calculated based on the number of these pixels. Then, the environmentally adaptive machine vision device 120 determines whether the product is good or bad by comparing the calculated score of the received product image with a reference threshold (S320).

In addition, the environmentally adaptive machine vision device 120 can transmit results related to the good or bad quality of the product to managers at the production line, etc. (S330). If there is a monitor on the production line, the monitor can be used to notify of non-good quality of a specific product.

Separately, the environmentally adaptive machine vision device 120 can calculate the ratio of good and bad products. More specifically, the environmentally adaptive machine vision device 120 calculates the overall NG ratio and also calculates the current NG ratio (S350, S360). The overall NG ratio measures the number of OK/NG and calculates the ratio of NG, that is, RN, based on the inference results. This can be expressed as <Equation 1>.

<Equation 1> may be expressed in the form RN = NG_all/(OK_all + NG_all).

Additionally, the current NG ratio measures the recent 1k inference results to calculate the number of OK/NG and the ratio of NG (=Rn). It can be expressed as <Equation 2>.

<Equation 2> can also be expressed as Rn = NG_1k / 1000.

Based on the total NG ratio and the current NG ratio, the environmentally adaptive machine vision device 120 determines whether the environment has changed (S370).

For example, if Rn / RN > 1.5, the environmentally adaptive machine vision device 120 may determine that NG occurs more than actually occurs due to environmental changes, and if Rn / RN < 0.66, NG may occur more than actually occurs due to environmental changes. It can be judged as a situation that occurs less frequently. Of course, standards for such situational judgments can be set through experiments, or it is possible to collect big data and analyze it through a deep learning program to obtain a standard value (e.g. 1.5, 0.66, etc.) based on the analysis results. It may be possible.

Furthermore, the environmentally adaptive machine vision device 120 performs threshold adjustment and result correction operations according to the determination of whether the environment has changed (S340). In other words, the environmentally adaptive machine vision device 120 calculates the threshold value Rn = RN using the score values of the recent 1k inference results, and updates the recent 1k inference results by applying a new threshold. For example, after sorting the most recent 1k inference results in order of score value, the score of the RN × 1000th image can be changed to the threshold.

In addition to the above, the imaging device 100 and the environmentally adaptive machine vision device 120 of FIG. 1 can perform various operations, and other details have been sufficiently explained previously, so these will be replaced.

FIG. 6 is a flowchart showing the operation process of the environmentally adaptive machine vision device of FIG. 1.

Referring to FIG. 6 together with FIG. 1 for convenience of explanation, the environmentally adaptive machine vision device 120 according to an embodiment of the present invention receives photographed data of the inspected product (S600). The captured image or video provided from the photographing device 100 may be compressed, that is, encoded, and transmitted, and the environmentally adaptive machine vision device 120 may perform a decoding operation to restore the compressed image, but according to the present invention, In the embodiment, it may mean pixel data that is photographed data in its restored state, that is, in the decoding state.

In addition, the environmentally adaptive machine vision device 120 analyzes the received shooting data and detects changes in the ratio of the product's non-good state (NG: No Good) based on the analysis results, and according to the detected result, the inspected product Changes in the shooting environment are determined and the analysis results are corrected based on the judgment results (S610).

Here, correcting can have various meanings. For example, in addition to changing the preset threshold, the threshold is fixed, the pixel data is corrected by the correction value, a score is calculated, and this is combined with the reference threshold. The purpose is to distinguish OK or NG products by comparison. However, in an embodiment of the present invention, the height of the preset threshold can be adjusted, and in the process, if correction is needed for data before adjustment, a correction operation can be performed.

In addition to the above, the environmentally adaptive machine vision device 120 of FIG. 1 can perform various operations, and other details have been sufficiently explained above, so these will be replaced.

Meanwhile, even though all the components constituting the embodiment of the present invention are described as being combined or operating in combination, the present invention is not necessarily limited to this embodiment. That is, as long as it is within the scope of the purpose of the present invention, all of the components may be operated by selectively combining one or more of them. In addition, although all of the components may be implemented as a single independent hardware, a program module in which some or all of the components are selectively combined to perform some or all of the combined functions in one or more pieces of hardware. It may also be implemented as a computer program having. The codes and code segments that make up the computer program can be easily deduced by a person skilled in the art of the present invention. Such computer programs can be stored in non-transitory computer readable media and read and executed by a computer, thereby implementing embodiments of the present invention.

Here, a non-transitory readable recording medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short period of time, such as a register, cache, or memory. . Specifically, the above-described programs may be stored and provided on non-transitory readable recording media such as CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM, etc.

In the above, preferred embodiments of the present invention have been shown and described, but the present invention is not limited to the specific embodiments described above, and may be used in the technical field to which the invention pertains without departing from the gist of the invention as claimed in the claims. Of course, various modifications can be made by those skilled in the art, and these modifications should not be understood individually from the technical idea or perspective of the present invention.

100: photographing device 110: communication network
120: Environmentally adaptive machine vision device 200: Communication interface unit
210: Control unit 220: Environmentally adaptive machine vision unit
230: storage unit

Claims (10)

A communication interface unit that receives photographed data of the inspected product; and
The received shooting data is analyzed to detect changes in the ratio of non-good condition (NG) of the product based on the analysis results, and the change in the shooting environment of the inspected product is determined based on the detection results. A control unit that corrects the analysis results;
An environmentally adaptable machine vision device including: According to paragraph 1,
The control unit determines the unhealthy state based on a preset threshold, and automatically adjusts the preset threshold based on the determination result to correct the analysis result. According to paragraph 2,
The control unit automatically sets the threshold value higher than the preset threshold value when the ratio of the unhealthy state increases, and automatically sets the threshold value lower than the preset threshold value when the ratio of the unhealthy state decreases. Adaptive machine vision device. According to paragraph 2,
The control unit is an environmentally adaptive machine vision device that automatically adjusts a preset threshold by determining changes in the environment in consideration of changes in the intensity or location of lighting in the space where product inspection is performed or changes in the location of the product as a result of the change in the shooting environment. According to paragraph 1,
The control unit calculates an analysis result of the received photographed data into a score and compares the calculated score with a preset threshold to determine the unhealthy state. Receiving, by a communication interface unit, photographic data of an inspected product; and
The control unit analyzes the received shooting data and detects changes in the rate of non-good condition (NG) of the product based on the analysis results, and determines changes in the shooting environment of the inspected product according to the detected results to provide a decision result. Correcting the analysis results based on
A method of operating an environmentally adaptive machine vision device including: According to clause 6,
The correction step is,
A method of driving an environment-adaptive machine vision device that determines the unhealthy state based on a preset threshold and corrects the analysis result by automatically adjusting the preset threshold based on the determination result. In clause 7,
The correction step is,
Environmentally adaptive machine vision that automatically sets the threshold higher than the preset threshold when the ratio of unhealthy states increases, and automatically sets the threshold lower than the preset threshold when the ratio of unhealthy states decreases. How to operate the device. In clause 7,
The correction step is,
A method of driving an environmentally adaptive machine vision device that determines changes in the environment and automatically adjusts a preset threshold by taking into account changes in the intensity or location of lighting in the space where product inspection is performed or changes in the location of the product due to changes in the shooting environment. According to clause 6,
The correction step is,
A method of driving an environmentally adaptive machine vision device that calculates the analysis result of the received shooting data into a score and compares the calculated score with a preset threshold to determine the unhealthy state.

Images