KR20240047550A - Atomic content ratio analysis method and device using vision camera - Google Patents

Abstract

The present invention includes the steps of photographing an object to be analyzed for atomic content moving through a conveyor unit through a photon input sensor and obtaining image information, and a control unit using the image information through pre-trained artificial intelligence to determine the object to be analyzed for atomic content. In the step of calculating the weight correction coefficient, the control unit calculates the predicted weight of the object to be analyzed for atomic content using a first calculation formula, radiates neutrons to the object to be analyzed for atomic content by a neutron generator, and uses a gamma-ray detector. Atomic content ratio analysis method using a vision camera, comprising the step of detecting gamma rays emitted by the object to be analyzed for atomic content and calculating the atomic content ratio of the object to be analyzed by the control unit using a second calculation formula. provides.

Description

Atomic content ratio analysis method and device using vision camera}

The present invention relates to an atomic content ratio analysis method and device using a vision camera, and more preferably, to perform weight correction using image information acquired through a vision camera and analyze the atomic content ratio of the object to be analyzed. This relates to the atomic content ratio analysis method and device using a vision camera.

The analysis of substances or materials is of great importance in many industrial sectors, especially with regard to their elemental composition, especially with regard to hazardous substances or waste or recycled materials or raw materials, or in the quality control of semi-finished products or industrial products. One of the previously performed analytical methods is the so-called multi-element analysis, in which the individual elements of a sample are determined without the need to know the exact composition of the sample in advance.

Multi-element analysis may be implemented by neutron activation or alternatively by, for example, X-ray fluorescence spectrometry or mass spectrometry. Previously, multi-element analysis by neutron activation was implemented through irradiation according to specific time specifications. In the case of pulsed neutron irradiation, it has been shown that meaningful measurement results can be guaranteed if, after a certain time window, prompt gamma radiation is evaluated according to the method of pulsed irradiation. The time window for detection of gamma rays begins after the waiting time after the end of each neutron pulse and ends before the next neutron pulse is emitted.

It is a nuclear reaction chemical detection technology that uses various types of neutron nuclear reaction techniques by converting fast neutrons (~2.5 MeV) from the D-D nuclear reaction of a neutron generator into thermal neutrons, the so-called Prompt Gamma-ray Neutron Activation Analysis. , hereinafter PGNAA) is widely known. In addition, Pulsed Fast Thermal Neutron Analysis (PFTNA) is a component analyzer of baggage for contraband such as explosives and drugs.

Component analyzers using PGNAA or PFTNA are currently used in mining and security (explosives, drugs). In order to increase the accuracy of analysis of the content ratio of each element, the weight of the object passing through the analyzer must be known in real time, and the total weight of each element must be calculated. Significant accuracy can be achieved only by detecting how many gamma rays are measured. Currently, weight measurement uses methods such as estimating the average amount of object transfer through conveyor speed and measurement using a scale-combined conveyor.

Conventional component analysis methods using PGNAA or PFTNA have a component ratio measurement error rate of 5 to 10%, which has the problem of not being applicable to industrial fields that require high accuracy.

Meanwhile, an artificial intelligence (AI) system is a computer system that implements human-level intelligence, and unlike existing rule-based smart systems, it is a system in which machines learn and make decisions on their own. As artificial intelligence systems are used, their recognition rates improve and they can more accurately understand user preferences, and existing rule-based smart systems are gradually being replaced by deep learning-based artificial intelligence systems.

Artificial intelligence technology consists of machine learning (deep learning) and element technologies using machine learning. Machine learning is an algorithmic technology that classifies/learns the characteristics of input data on its own, and elemental technology is a technology that uses machine learning algorithms such as deep learning to mimic the functions of the human brain such as cognition and judgment, including linguistic understanding and visual It consists of technical areas such as understanding, reasoning/prediction, knowledge expression, and motion control.

Patent Document 1: Republic of Korea Patent Publication No. 10-2009-0046805

The present invention was developed to solve the above problems, and atomic content ratio analysis using a vision camera performs weight correction using image information acquired through a vision camera and analyzes the atomic content ratio of the object to be analyzed. Provides a method and a device therefor.

In addition, we provide an atomic content ratio analysis method and device using a vision camera that combines component analysis using PGNAA or PFTNA and artificial intelligence data pre-learning technology.

The objects of the present invention are not limited to the objects mentioned above, and other objects of the present invention that are not mentioned can be understood by the following description and can be more clearly understood by examples of the present invention. Additionally, the objects of the present invention can be realized by means and combinations thereof as indicated in the claims.

The atomic content ratio analysis method using a vision camera to achieve the purpose of the present invention described above includes the following steps.

In one embodiment of the present invention, photographing an object to be analyzed for atomic content moving through a conveyor unit using a photon input sensor and obtaining image information; A control unit calculating a weight correction coefficient of the object to be analyzed for atomic content using the image information through pre-trained artificial intelligence; The control unit calculates the predicted weight of the object to be analyzed for atomic content using a first calculation formula, radiates neutrons to the object to be analyzed for atomic content by a neutron generator, and radiates neutrons from the object to be analyzed for atomic content by a gamma ray detector. detecting gamma rays; and calculating, by the control unit, the atomic content ratio of the object to be analyzed for atomic content using a second calculation formula; Provides an atomic content ratio analysis method using a vision camera including.

In addition, the first calculation formula provides an atomic content ratio analysis method using a vision camera, wherein V*a*C = χ (kg/s).

In _addition , the second calculation formula _{is T} Provides a content ratio analysis method.

In addition, the photon input sensor provides an atomic content ratio analysis method using a vision camera, wherein the photon input sensor is a type selected from the group consisting of a vision camera, radar, lidar, and ultrasonic sensor.

In addition, the a value provides an atomic content ratio analysis method using a vision camera in which the a value is calculated using an average object transfer amount estimation method through conveyor speed.

In addition, the C value is calculated to correct the a value through pre-learned artificial intelligence by transmitting image information data collected in real time through the photon input sensor to the control unit. An atomic content ratio analysis method using a vision camera. provides.

The atomic content ratio analysis device using a vision camera to achieve the purpose of the present invention described above includes the following steps.

In one embodiment of the present invention, a conveyor unit configured to move an object to be analyzed for atomic content; A photon input sensor formed on the conveyor unit and acquiring image information by photographing the object to be analyzed for atomic content moving through the conveyor unit; a neutron generator formed on the conveyor unit and radiating neutrons to the object to be analyzed for atomic content; A gamma ray detector configured to detect gamma rays emitted by the object to be analyzed for atomic content; Collect the image information, calculate the weight information of the object for atomic content analysis, calculate the weight correction coefficient of the object for atomic content analysis through pre-learned artificial intelligence, and use the first calculation formula to calculate the weight of the object for atomic content analysis. Calculate the predicted weight, irradiate neutrons to the object to be analyzed for atomic content through the neutron generator, receive gamma rays detected through the gamma-ray detector, and calculate the atomic content of the object to be analyzed for atomic content using a second calculation formula. A control unit that calculates the ratio; Provides an atomic content ratio analysis device using a vision camera configured to include.

In addition, the first calculation formula provides an atomic content ratio analysis device using a vision camera, wherein V*a*C = χ (kg/s).

In _addition , the second calculation formula _{is T} A content ratio analysis device is provided.

In addition, the photon input sensor provides an atomic content ratio analysis device using a vision camera, wherein the photon input sensor is a type selected from the group consisting of a vision camera, radar, lidar, and ultrasonic sensor.

In addition, an atomic content ratio analysis device using a vision camera is provided where the a value is calculated using an average object transfer amount estimation method through conveyor speed.

In addition, the C value is calculated to correct the a value through pre-learned artificial intelligence by transmitting image information data collected in real time through the photon input sensor to the control unit. Atomic content ratio analysis device using a vision camera provides.

The present invention can achieve the following effects by combining the above-mentioned embodiment with the configuration, combination, and use relationship described below.

It has the effect of improving the accuracy of component ratio measurement by performing weight correction using image information acquired through a vision camera and analyzing the atomic content ratio of the object to be analyzed.

By combining component analysis using PGNAA or PFTNA and artificial intelligence data pre-learning technology, prediction error is reduced through deep learning analysis compared to existing analysis-based performance analysis, an effect that can be practically applied to industrial fields that require accuracy. has

Figure 1 shows an overview of the entire process of the atomic content ratio analysis method using a vision camera as an embodiment of the present invention.
Figure 2 shows an atomic content ratio analysis device using a vision camera as an embodiment of the present invention.
Figure 3 shows the configuration of an atomic content ratio analysis device using a vision camera as an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the attached drawings. Embodiments of the present invention may be modified in various forms, and the scope of the present invention should not be construed as being limited to the following embodiments. This example is provided to more completely explain the present invention to those skilled in the art.

Additionally, terms such as "... unit" and "... unit" described in the specification refer to a unit that processes at least one function or operation, and may be implemented as software or hardware.

The identification code for each step is used for convenience of explanation. The identification code does not explain the order of each step, and each step may be performed differently from the specified order unless a specific order is clearly stated in the context. there is.

In addition, when a part described in the specification is said to be "connected" to another part, this includes not only direct connection but also indirect connection, and indirect connection refers to connection through a wireless communication network. Includes.

Additionally, terms such as first and second described in the specification are used to distinguish one component from another component, and the components are not limited by the above-mentioned terms.

Figure 1 shows an overview of the entire process of the atomic content ratio analysis method using a vision camera as an embodiment of the present invention.

Referring to FIG. 1, the atomic content ratio analysis method using a vision camera according to an embodiment of the present invention photographs the object to be analyzed for atomic content moving through the conveyor unit 100 through the photon input sensor 200. It may include acquiring image information.

The photon input sensor 200 may be configured to photograph an object to be analyzed for atomic content moving through the conveyor unit 100. More preferably, the photon input sensor 200 may be configured to acquire image information by photographing the object to be analyzed for atomic content moving at a constant speed through the conveyor unit 100 in real time. In one embodiment, the photon input sensor 200 may obtain image information by photographing an object to be analyzed for atomic content in 1 second units.

The photon input sensor 200 may be one type selected from the group consisting of a vision camera, radar, lidar, and ultrasonic sensor. More preferably, the photon input sensor 200 may be a vision camera applicable to conveyor lines and operating even at night.

The photon input sensor 200 may be connected to the artificial intelligence device and the control unit 500 and configured to transmit the acquired image information to the artificial intelligence device. Additionally, image information acquired through the photon input sensor 200 may be transmitted to the control unit 500. The present invention includes the steps of the control unit 500 calculating the weight correction coefficient of the object to be analyzed for atomic content using image information through pre-trained artificial intelligence and calculating the predicted weight of the object to be analyzed for atomic content using the first calculation formula. It may include steps.

The first calculation formula may be V*a*C = χ (kg/s).

here,

- V is the conveyor speed (m/s),

- a is the weight information (kg/m) of the object subject to atomic content analysis per 1 m of conveyor,

- C is the weight correction coefficient derived from the vision camera analysis results,

- χ is the predicted weight (kg/s) of the object subject to atomic content analysis per unit time.

The a value can be calculated using the method of estimating the average object transport amount through conveyor speed. The C value can be calculated by transmitting image information data collected in real time through the photon input sensor 200 to the control unit 500 and applying the C value to the a value to calculate the χ value.

More specifically, in the step (S100) of calculating the predicted weight of the object to be analyzed for atomic content using the first calculation formula, the a value can be corrected through the value of C using real-time image information of the object to be analyzed for atomic content. . The value calculated through the first calculation formula can be included as a factor in the second calculation formula described later, thereby improving the analysis accuracy of the object to be analyzed for atomic content.

In the step where the control unit 500 calculates the weight correction coefficient of the object to be analyzed for atomic content using image information through pre-learned artificial intelligence, the control unit 500 performs the pre-learning in real time through the photon input sensor 200. This can be done with big data by dividing the collected image information data into parts with and without the target object. More specifically, dictionary learning can be performed with big data in which the control unit 500 classifies image information data collected in real time by the presence or absence of a target object within the captured image. The presence or absence of the target object within the captured image can be determined, for example, by the size of the voids between particles of the target object. The void between the particles of the target object may mean a place where the conveyor belt is visible because there is no target object inside the captured image taken by the vision camera. The gap may become larger when large particle objects are concentrated on the conveyor belt, and may become smaller when small particle objects are concentrated than when large particles are concentrated on the conveyor belt. Additionally, the voids can become smaller as the objects of large and small particles are evenly packed.

The control unit 500 can calculate the C value through pre-trained artificial intelligence by matching the weight information of the object to be analyzed for atomic content per 1 m of conveyor collected in real time and the classified image information data. As an example, the control unit 500 may calculate a weight correction coefficient for the size of the air gap by matching the classified image information data with the weight information of the object to be analyzed.

In one embodiment, the control unit 500 may classify image information data collected in real time by whether or not iron ore is located within the captured image. More specifically, voids between particles of iron ore can indicate where the conveyor belt is visible due to the absence of iron ore inside the captured images taken by the vision camera.

The image information data classified in this way can be transmitted to an artificial intelligence device and pre-trained using the correlation with the atomic content per meter of conveyor and the weight information of the object to be analyzed as learning data. As an example, the control unit 500 may classify and process data for such prior learning, transmit a plurality of data to an artificial intelligence device, and calculate a weight correction coefficient using big data previously learned by the artificial intelligence device.

For example, when there are many voids, the estimated cargo volume (m.kg/s) is measured to be larger than the actual cargo volume (m.kg/s), and in this case, the weight correction coefficient may have a value of less than 1. Therefore, the estimated cargo volume (m.kg/s) corrected through the first calculation formula is smaller than the uncorrected estimated cargo volume (m.kg/s) and can be close to the actual cargo volume (m.kg/s). . In addition, if there are no voids and there are many overlapping areas of iron ore on the conveyor belt, the estimated cargo volume (m.kg/s) is measured to be smaller than the actual cargo volume (m.kg/s), and in this case, the weight correction coefficient is a value exceeding 1. You can have Therefore, the estimated cargo volume (m.kg/s) corrected through the first calculation formula is larger than the uncorrected estimated cargo volume (m.kg/s) and can be close to the actual cargo volume (m.kg/s). Through this, the error rate between actual cargo volume (m.kg/s) and estimated cargo volume (m.kg/s) due to the particle size and porosity of iron ore can be reduced, and the accuracy of analysis of iron (Fe) content contained in iron ore can be improved. You can do it.

The present invention includes a step (S200) in which the control unit 500 irradiates neutrons to the object to be analyzed for atomic content by the neutron generator 300 and detects gamma rays emitted by the object to be analyzed for atomic content by the gamma ray detector 400. It can be included. More specifically, when the control unit 500 transmits a neutron irradiation signal to the neutron generator 300, the neutron generator 300 can irradiate neutrons to the object to be analyzed for atomic content. When gamma rays are emitted by irradiating neutrons to an object to be analyzed for atomic content, the control unit 500 can detect the gamma rays using the gamma ray detector 400.

As an example, the control unit 500 may detect gamma rays to calculate the atomic content ratio of iron (Fe) atoms when the object to be analyzed for atomic content is iron ore. When the control unit 500 irradiates neutrons, the irradiated neutrons hit the nucleus of the iron atom, and the iron atom becomes excited and emits gamma rays to return to low energy. At this time, when the gamma ray emitted from the front scintillator of the gamma ray detector 400 hits, the energy band of the iron atom is lowered and converted to a level of light that can react with the optical sensor and amplified, and the optical sensor converts the received light into an electrical signal. It can be transmitted to the control unit 500.

The present invention may include a step (S300) in which the control unit 500 calculates the atomic content ratio of the object to be analyzed for atomic content using a second calculation formula. More specifically, the control unit 500, which receives the signal of the number of gamma rays of the object to be analyzed for atomic content detected through the gamma ray detector 400, can calculate the atomic content ratio of the object to be analyzed for atomic content using the second calculation formula. .

_The second calculation formula may _{be T}

here,

- _T

- _α

- α _all is the number of gamma rays of all atoms of the object subject to atomic content analysis,

- Ar(X) is the atomic weight of the X atom of the object to be analyzed for atomic content,

- Ar(all) is the atomic weight of all atoms of the object subject to atomic content analysis,

- Z(X) is the neutron reaction rate correction coefficient of the X atom of the object to be analyzed for atomic content,

- Z(all) is the average correction coefficient for the neutron reaction rate of all atoms of the object to be analyzed for atomic content,

- χ is the predicted weight of the object subject to atomic content analysis per unit time (kg/s).

The control unit 500 can derive the corrected χ value through the first calculation formula, and calculate a more accurate _T The control unit 500 is an Electronic Control Unit (ECU), a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), an Application Processor (AP), an Application Processor (AP), or an application processor in the technical field of the present invention. It may be configured to include at least one processor of any type well known to the. Additionally, the control unit 500 may be comprised of a combination of software and hardware capable of performing operations on at least one application or program for executing methods according to embodiments of the present invention.

According to an embodiment of the present invention, the accuracy of the atomic content ratio of the object subject to atomic content analysis can be improved compared to the comparative example below. Hereinafter, improvement in atomic content analysis accuracy will be explained through comparative examples and examples when the object to be analyzed for atomic content is iron ore.

<Comparison example>

Time (S) 38276 38277 38278 38279 38280 38281 38282 Actual cargo volume (m.kg/s) 24.2 24.7 23.1 23.4 21.8 22.7 25.1 Estimated cargo volume (m.kg/s) 25.0 25.0 25.0 25.0 25.0 25.0 25.0

<Example>

Time (S) 38276 38277 38278 38279 38280 38281 38282 Actual cargo volume (m.kg/s) 24.2 24.7 23.1 23.4 21.8 22.7 25.1 Corrected estimated cargo volume (m.kg/s) 24.6 24.9 22.9 23.4 21.7 22.8 25.0

If the 24-hour average iron ore volume is 25.0 (m.kg/s) and the unit time is 1 second, in the comparative example, the actual iron ore volume for the section is 23.6 (m.kg/s) and the estimated iron ore volume for the section is 25.0 (m.kg/s). s), the accuracy of the iron content (%) analysis of iron ore decreases due to the error between the actual iron ore volume in the section and the estimated iron ore volume in the section.

On the other hand, in the embodiment under the same conditions as the comparative example, the actual iron ore volume for the section is 23.6 (m.kg/s) and the estimated iron ore volume for the corrected section is 23.6 (m.kg/s), so the actual iron ore volume for the section and the estimated section are By minimizing the error between iron ore shipments, the accuracy of iron (Fe) content (%) analysis of iron ore can be improved. Figure 2 shows an atomic content ratio analysis device using a vision camera as an example of the present invention, and Figure 3 is a configuration diagram of an atomic content ratio analysis device using a vision camera as an embodiment of the present invention. It shows.

Referring to Figures 2 and 3, the atomic content ratio analysis device using a vision camera according to an embodiment of the present invention includes a conveyor unit 100, a photon input sensor 200, a neutron generator 300, and a gamma ray detector. It may include 400 and a control unit 500. The conveyor unit 100 may be configured to install a photon input sensor 200, a neutron generator 300, and a gamma-ray detector 400, and the control unit 500 may include a photon input sensor 200 and a neutron generator. It may be connected to 300 and the gamma-ray detector 400 and configured to transmit and receive signals.

The conveyor unit 100 may be configured to move the object to be analyzed for atomic content. More preferably, the conveyor unit 100 may include a conveyor belt that moves at a constant speed and a conveyor housing. The photon input sensor 200 may be formed on the conveyor unit 100 and configured to obtain image information by photographing an object to be analyzed for atomic content moving through the conveyor unit 100.

As an example, when an object to be analyzed for atomic content moves on the top of a conveyor belt, a vision camera fixed to the conveyor housing may be configured to photograph the object in real time. More preferably, the vision camera may be configured to acquire image information by photographing the object to be analyzed for atomic content in 1 second units.

The neutron generator 300 may be formed in the conveyor unit 100 and configured to radiate neutrons to the object to be analyzed for atomic content. More preferably, as shown in FIG. 3, the neutron generator 300 may be fixedly installed on the conveyor housing at a distance of 1 m from the vision camera. The gamma ray detector 400 may be configured to detect gamma rays emitted by an object to be analyzed for atomic content. More preferably, the gamma ray detector 400 may be fixedly installed on the conveyor housing and spaced apart from the neutron generator 300.

In one embodiment, when the object to be analyzed for atomic content is iron ore, the neutron generator 300 emits high-energy fast neutrons, and the fast neutrons react with the iron ore, losing energy at a rapid rate and entering the thermal neutron region. (thermal energy level). At this time, after the fast neutrons are emitted, they interact with iron ore through inelastic scattering until the energy drops below about 1 MeV. In other words, the fast neutron collides with the nucleus of the iron (Fe) atomic material in iron ore and excites the atomic nucleus, and as the excited nucleus is restored to its original state, inelastic scattering gamma rays are emitted. On the other hand, when the neutron loses more energy and falls to the level of 0.025 eV and becomes a thermal neutron, a neutron capture phenomenon occurs in which the thermal neutron is absorbed by the nucleus of the iron atom in the iron ore. In the process where a neutron is captured in an atomic nucleus, the nucleus is excited and then restored, emitting a capture gamma ray. Inelastic scattering gamma rays and captured gamma rays emitted from iron ore due to interaction with neutrons irradiated by the neutron generator 300 can be measured by a gamma ray detector.

The shield shown in FIG. 3 is formed between the neutron generator 300 and the gamma-ray detector 400, and may also be formed outside the neutron generator 300 and the gamma-ray detector 400. The shield may be configured to prevent neutrons emitted from the neutron generator 300 from directly entering the gamma-ray detector 400. The shielding body may be composed of a neutron absorbing material and a neutron scattering material. The neutron absorbing material may be boron, natural boron, boron-containing polyene, or tylene, and the neutron scattering material may include tungsten.

The control unit 500 may be configured to collect image information, calculate weight information of the object to be analyzed for atomic content, and calculate the weight correction coefficient of the object to be analyzed for atomic content through pre-trained artificial intelligence. In addition, the control unit 500 calculates the predicted weight of the object to be analyzed for atomic content using the first calculation formula, irradiates the object to be analyzed for atomic content with neutrons through the neutron generator 300, and radiates neutrons to the object to be analyzed for atomic content through the gamma ray detector 400. It may be configured to receive the detected gamma rays and calculate the atomic content ratio of the object to be analyzed for atomic content using a second calculation formula.

More specifically, the control unit 500 is connected to the photon input sensor 200 and an artificial intelligence (AI) device to collect image information obtained through the photon input sensor 200 to determine the weight correction coefficient of the object for atomic content analysis. can be calculated.

The control unit 500 can classify the image information data collected in real time through the photon input sensor 200 into parts with and without the target object, and pre-learning allows the AI device to receive these data from the control unit 500. It can be received in multiple numbers and performed as big data.

More specifically, the control unit 500 can classify image information data collected in real time into the presence or absence of a target object within the captured image and transmit it to the AI device. When the control unit 500 determines whether a target object is located within a captured image, the decision may be made based on the size of the voids between particles of the target object. At this time, the control unit 500 may recognize the voids between the particles of the object as a place where the conveyor belt is visible because there is no object inside the captured image taken by the vision camera. The control unit 500 may determine that the void is large when large particle objects are crowded on the conveyor belt, and may determine that the void is smaller when small particle objects are densely packed than when large particles are densely packed. Additionally, the control unit 500 may determine that the void is smaller as the objects of large and small particles are evenly packed.

The control unit 500 can calculate the C value through pre-trained artificial intelligence by matching the weight information of the object to be analyzed for atomic content per 1 m of conveyor collected in real time and the classified image information data. In one embodiment, the vision camera captures an image of iron ore in response to the moving speed of the conveyor belt, and the control unit 500 can classify the image information data collected in real time into a part with iron ore and a part without iron ore. The control unit 500 transmits image information data classified into parts with iron ore and parts without iron ore to the AI device so that the artificial intelligence device pre-learns the correlation between the atomic content per 1 m of conveyor and the weight information of the object to be analyzed as learning data. can do. The control unit 500 can classify and process the data for pre-learning, transmit a plurality of data to the AI device, and calculate the weight correction coefficient using big data pre-learned by the AI device.

For example, if the estimated cargo volume (m.kg/s) is measured to be larger than the actual cargo volume (m.kg/s) due to a large number of voids, the control unit 500 may calculate the weight correction coefficient to a value of less than 1. . In addition, the control unit 500 sets the weight correction coefficient by exceeding 1 when the estimated cargo volume (m.kg/s) is measured to be smaller than the actual cargo volume (m.kg/s) due to there being no air gap and many overlapping areas of iron ore on the conveyor belt. It can be calculated as the value of .

The control unit 500 can use the calculated weight correction coefficient to calculate the predicted weight of iron ore from the first calculation formula and calculate the atomic content ratio of iron atoms of the iron ore through the second calculation formula. Through this, the control unit 500 can reduce the error rate between the actual cargo volume (m.kg/s) and the estimated cargo volume (m.kg/s) due to the particle size and porosity of the iron ore, and the iron (Fe) content contained in the iron ore. The accuracy of analysis can be improved.

The AI device may include an electronic device including an AI module capable of performing AI processing or a server including an AI module. Additionally, the AI device may be included as part of at least a portion of an electronic device and may be equipped to perform at least part of AI processing. The AI device may include an AI processor, memory, and/or communication unit. AI devices are computing devices that can learn neural networks and can be implemented in various electronic devices such as servers, desktop PCs, laptop PCs, and tablet PCs.

AI processors can learn neural networks using programs stored in memory. In particular, the AI processor can learn a neural network to recognize image information data for predicting the atomic content of the object subject to atomic content analysis. The AI processor may be a general-purpose processor (e.g., CPU), but may be an AI-specific processor (e.g., GPU) for artificial intelligence learning. Memory can store various programs and data necessary for the operation of AI devices. The memory is accessed by the AI processor, and data reading/recording/modification/delete/update, etc. can be performed by the AI processor. Additionally, the memory may store a neural network model (eg, deep learning model) generated through a learning algorithm for classifying/recognizing image information data according to an embodiment of the present specification. The communication unit may transmit the results of AI processing by the AI processor to the control unit 500. The control unit 500 collects image information in real time from the photon input sensor 200 and transmits the collected image information data to the AI device for pre-learning to calculate the weight correction coefficient of the object to be analyzed for atomic content.

In summary, the present invention performs weight correction using image information acquired through a vision camera and analyzes the atomic content ratio of the object to be analyzed, thereby improving the accuracy of component ratio measurement. Atomic content ratio analysis using a vision camera Provides a method and a device therefor.

The above detailed description is illustrative of the present invention. Additionally, the foregoing is intended to illustrate preferred embodiments of the present invention, and the present invention can be used in various other combinations, modifications, and environments. That is, changes or modifications can be made within the scope of the concept of the invention disclosed in this specification, a scope equivalent to the disclosed content, and/or within the scope of technology or knowledge in the art. The described embodiments illustrate the best state for implementing the technical idea of the present invention, and various changes required for specific application fields and uses of the present invention are also possible. Accordingly, the detailed description of the invention above is not intended to limit the invention to the disclosed embodiments. Additionally, the appended claims should be construed to include other embodiments as well.

100: conveyor unit
200: Photon input sensor
300: Neutron generator
400: Gamma ray detector
500: Control unit

Claims (12)

Photographing an object to be analyzed for atomic content moving through a conveyor unit using a photon input sensor and acquiring image information;
A control unit calculating a weight correction coefficient of the object to be analyzed for atomic content using the image information through pre-trained artificial intelligence;
The control unit calculates the predicted weight of the object to be analyzed for atomic content using a first calculation formula, radiates neutrons to the object to be analyzed for atomic content by a neutron generator, and radiates neutrons from the object to be analyzed for atomic content by a gamma ray detector. detecting gamma rays; and
A control unit calculating the atomic content ratio of the object to be analyzed for atomic content using a second calculation formula; containing
Atomic content ratio analysis method using a vision camera.
According to claim 1,
The first calculation formula is,
Characterized in that V*a*C = χ (kg/s)
Atomic content ratio analysis method using a vision camera.
(V: Conveyor speed (m/s), a: Weight information of the object for atomic content analysis per 1 m (kg/m), C: Weight correction coefficient derived from vision camera analysis results, χ: Corrected object for atomic content analysis Transfer amount (kg/s))
According to claim 1,
The second calculation formula is,
_{Characterized} in _{that T}
Atomic content ratio analysis method using a vision camera.
_{( T ​} Atomic weight of X atoms of object subject to content analysis, Ar(all): Atomic weight of all atoms of object subject to atomic content analysis, Z(X): Neutron reaction rate correction coefficient of Average correction coefficient for the neutron reaction rate of all atoms of the object, χ: predicted weight of the object analyzed for atomic content per unit time (kg/s))
According to claim 1,
The photon input sensor is characterized in that it is a type selected from the group consisting of a vision camera, radar, lidar, and ultrasonic sensor.
Atomic content ratio analysis method using a vision camera.
According to claim 2,
The a value is calculated using the average object transfer amount estimation method through conveyor speed.
Atomic content ratio analysis method using a vision camera.
According to claim 2,
The C value is calculated to correct the a value through pre-trained artificial intelligence by transmitting image information data collected in real time through the photon input sensor to the control unit.
Atomic content ratio analysis method using a vision camera.
A conveyor unit configured to move the object to be analyzed for atomic content;
A photon input sensor formed on the conveyor unit and acquiring image information by photographing the object to be analyzed for atomic content moving through the conveyor unit;
a neutron generator formed on the conveyor unit and radiating neutrons to the object to be analyzed for atomic content;
A gamma ray detector configured to detect gamma rays emitted by the object to be analyzed for atomic content;
Collect the image information, calculate the weight information of the object to be analyzed for atomic content, and calculate the weight correction coefficient of the object to be analyzed for atomic content through pre-trained artificial intelligence,
Calculate the predicted weight of the object to be analyzed for atomic content using a first calculation formula, radiate neutrons to the object to be analyzed for atomic content through the neutron generator, receive gamma rays detected through the gamma ray detector, and calculate the predicted weight of the object to be analyzed for atomic content using a second calculation formula. a control unit that calculates the atomic content ratio of the object to be analyzed for atomic content using; configured to include
Atomic content ratio analysis device using a vision camera.
According to claim 7,
The first calculation formula is,
Characterized in that V*a*C = χ (kg/s)
Atomic content ratio analysis device using a vision camera.
(V: Conveyor speed (m/s), a: Weight information of the object for atomic content analysis per 1 m (kg/m), C: Weight correction coefficient derived from vision camera analysis results, χ: Corrected object for atomic content analysis Transfer amount (kg/s))
According to claim 7,
The second calculation formula is,
_{Characterized} in _{that T}
Atomic content ratio analysis device using a vision camera.
_{( T ​} Atomic weight of X atoms of object subject to content analysis, Ar(all): Atomic weight of all atoms of object subject to atomic content analysis, Z(X): Neutron reaction rate correction coefficient of Average correction coefficient for the neutron reaction rate of all atoms of the object, χ: predicted weight of the object analyzed for atomic content per unit time (kg/s))
According to claim 7,
The photon input sensor is characterized in that it is a type selected from the group consisting of a vision camera, radar, lidar, and ultrasonic sensor.
Atomic content ratio analysis device using a vision camera.
According to claim 8,
The a value is calculated using the average object transfer amount estimation method through conveyor speed.
Atomic content ratio analysis device using a vision camera.
According to claim 8,
The C value is calculated to correct the a value through pre-trained artificial intelligence by transmitting image information data collected in real time through the photon input sensor to the control unit.
Atomic content ratio analysis device using a vision camera.

Images