Information processing apparatus

Abstract

The present technology relates to an information processing apparatus capable of more easily determining a maintenance time by using event data. The information processing apparatus includes a state estimation unit that estimates a state of an abrasive stone by using event data provided from an event sensor that outputs a temporal change in an electric signal obtained by photoelectrically converting a light signal as the event data, and outputs a result of the estimation. The present technology can be applied to, for example, an information processing system that notifies a maintenance time of an abrasive stone of a machine tool, and the like.

Description

Technical Field

[0001] This technology relates to information processing apparatus, and more specifically, to an information processing apparatus that can more easily determine maintenance times using event data. Background Technology

[0002] Patent Document 1 discloses a maintenance support device that generates a learning model by performing machine learning using a learning dataset, in which the actual surface roughness measured by an external measuring device is the target variable, the measurement data measured by an internal measuring device is the descriptive variable, and the device performs a process to support the maintenance of machine tools based on the measurement data obtained by the internal measuring device (such as a non-contact displacement sensor).

[0003] Reference List

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2020-114615 Summary of the Invention

[0006] The problem to be solved by the present invention

[0007] The goal is to make it easier to determine the maintenance schedule for machine tools.

[0008] This technology was developed in light of this situation and makes it easier to determine maintenance times using event data.

[0009] Solution to the problem

[0010] The information processing apparatus according to an aspect of the present technology includes a state estimation unit that estimates the state of the grinding stone by using event data provided from an event sensor and outputs the estimation result. The event sensor outputs the time change of an electrical signal obtained by photoelectric conversion of an optical signal as event data.

[0011] According to this technology, the time variation of an electrical signal obtained by photoelectric conversion of an optical signal is output as event data, the state of the grinding stone is estimated by using the event data, and the estimated result is output.

[0012] The information processing device can be a standalone device or a module integrated into another device. Attached Figure Description

[0013] Figure 1 This is a block diagram illustrating a configuration example of a first embodiment of an information processing system applying this technology.

[0014] Figure 2 This is a diagram showing an instance of event data.

[0015] Figure 3 This is a diagram illustrating an example of a method for generating frame data from event data.

[0016] Figure 4 This is an illustration used to depict an event image capturing a falling spark.

[0017] Figure 5 This is a block diagram illustrating a detailed configuration example of an information processing device.

[0018] Figure 6 It is a diagram used to illustrate the relationship between measurement parameters and physical quantities.

[0019] Figure 7 It is a table showing the correlation between measurement parameters and physical quantities.

[0020] Figure 8 This is a flowchart illustrating the maintenance time determination process performed by the information processing system.

[0021] Figure 9 This is a flowchart used to illustrate the threshold update process.

[0022] Figure 10 This is a block diagram illustrating a configuration example of an EVS camera in a second embodiment of an information processing system applying the present technology.

[0023] Figure 11 This is a block diagram illustrating a detailed configuration example of the imaging element.

[0024] Figure 12 This is a block diagram illustrating a configuration example of an address event detection circuit.

[0025] Figure 13 This is a circuit showing the detailed configuration of the current-to-voltage conversion circuit, subtractor, and quantizer.

[0026] Figure 14 This is a diagram showing a more detailed configuration example of an address event detection circuit.

[0027] Figure 15 This is a circuit diagram showing another configuration example of a quantizer.

[0028] Figure 16 It is shown that in the adoption Figure 15 A more detailed circuit configuration example of the address event detection circuit in the case of a quantizer is shown in the diagram.

[0029] Figure 17 It is a block diagram illustrating an example of a computer's hardware configuration. Detailed Implementation

[0030] In the following description, embodiments for implementing the present technology (hereinafter referred to as implementations) will be described with reference to the accompanying drawings. Note that in the description and drawings, components having substantially the same function and configuration are indicated by the same reference numerals, and redundant descriptions are omitted. The descriptions will proceed in the following order.

[0031] 1. First embodiment of an information processing system

[0032] 2. Examples of event data

[0033] 3. Configuration Examples of Information Processing Devices

[0034] 4. Relationship between measurement parameters and physical quantities

[0035] 5. Flowchart for determining maintenance time

[0036] 6. Flowchart of threshold update process

[0037] 7. Second embodiment of the information processing system

[0038] 8. Conclusion

[0039] 9. Computer Configuration Examples

[0040] <1. First Embodiment of an Information Processing System>

[0041] Figure 1 A configuration example of a first embodiment of an information processing system applying this technology is shown.

[0042] Figure 1 The information processing system 1 includes an EVS camera 11, an information processing device 12, and a display 13; it is a system for estimating the condition of the grinding stone 22 of the machine tool 21 and informing the maintenance time.

[0043] Machine tool 21 is a machine tool that performs grinding operations on workpiece W, such as cylindrical grinding, internal surface grinding, and surface grinding, and is a so-called grinding machine. Machine tool 21 causes a grinding stone 22 to rotate at high speed to grind workpiece W. Coolant liquid 23 is supplied from an upper nozzle to the contact portion between the grinding stone 22 and workpiece W. During the grinding of workpiece W by the grinding stone 22, sparks 24 are generated from the contact portion between the grinding stone 22 and workpiece W. In addition to the sparks 24, coolant liquid 23 also drips.

[0044] The EVS camera 11 is a camera that includes an event sensor, which outputs the time-varying changes of an electrical signal obtained by photoelectric conversion of light signals as event data. This event sensor is also known as an event-based vision sensor (EVS). While a camera including a regular image sensor captures images synchronously with a vertical sync signal and outputs frame data of image data as a frame (picture) at the period of the vertical sync signal, the EVS camera 11 only outputs event data at the precise timing of an event. Therefore, the EVS camera 11 can be considered an asynchronous or address-controlled camera.

[0045] The EVS camera 11 is mounted such that its imaging range includes the workpiece W and the grinding stone 22 during the grinding process, detects changes in light (brightness) caused by sparks 24 generated during grinding and dripping coolant liquid 23 as events, and outputs the event data to the information processing device 12.

[0046] The information processing device 12 estimates the state of the grinding stone 22 based on event data output from the EVS camera 11. For example, the information processing device 12 determines whether the grinding stone 22 is clogged by processing the event data. If the information processing device 12 determines that the grinding stone 22 is clogged, the information processing device 12 outputs an alarm indicating that the grinding stone 22 is clogged. For the alarm, any method can be selected, such as outputting a sound, such as a buzzer, turning on a signal light, or displaying an alarm message. In this embodiment, the information processing device 12 causes the display 13 to display a message (text) such as "Clogged, maintenance required". In addition, the information processing device 12 uses the event data output from the EVS camera 11 to generate a display image and causes the display 13 to display the display image.

[0047] <2. Examples of Event Data>

[0048] Figure 2 An example of event data output by EVS camera 11 is shown.

[0049] For example, such as Figure 2 As shown, the EVS camera 11 outputs event data as an event, which includes the time t when the event occurs. i The coordinates (x, y) of the pixel where the event occurred. i y i And the polarity p of the brightness change i .

[0050] The time t of the event i It is a timestamp indicating the time when an event occurred, and is represented by, for example, the count value of a counter based on a predetermined clock signal in a sensor. In other words, the timestamp corresponding to the time when an event occurred is time information indicating the (relative) time of the event, as long as the interval between events is preserved as is.

[0051] polarity p i This indicates the direction of the brightness change when a brightness change (light intensity change) exceeding a predetermined threshold (hereinafter referred to as the event threshold) occurs as an event, and indicates whether the brightness change is in a positive direction (hereinafter also referred to as positive) or a negative direction (hereinafter also referred to as negative). For example, the polarity p of the event. i It is represented as "1" when it is positive and as "0" when it is negative.

[0052] exist Figure 2 In the event data, at time t of a certain event i and the time t of events adjacent to this event. i+1 The interval between them is not necessarily constant. That is, the time t of the event... i and time t i+1 These can be the same time or different times. However, assuming the event's time t... i+1 and time t i+1 There exists a given expression t i <=t i+1 The relationship indicated.

[0053] The EVS camera 11 outputs only the position coordinates, polarity, and time information of the pixels where brightness changes are detected. Because the EVS camera 11 only generates and outputs the net change (difference) in position coordinates, polarity, and time information, there is no redundancy in the data, and it has a high temporal resolution on the order of μs. Therefore, it can accurately capture instantaneously generated sparks 24, coolant liquid 23, etc.

[0054] Unlike image data (frame data) output in frame format during a frame period synchronized with the vertical sync signal, event data is output whenever an event occurs. Therefore, the event data cannot be displayed as an image on display 13, nor can it be input to an identifier (classifier) ​​for image processing, where display 13 displays the image corresponding to the frame data. To display the event data on display 13, it needs to be converted into frame data.

[0055] Figure 3 This is a diagram illustrating an example of a method for generating frame data from event data.

[0056] exist Figure 3 In the context of the event data, points are plotted as event data within a three-dimensional (time) space that includes the x-axis, y-axis, and time axis t. These points are plotted based on the event time t and the coordinates (x, y) of the pixels representing the event.

[0057] That is, assuming that the three-dimensional position (x, y, t) represented by the event's time t and the event's pixel (x, y) included in the event data is called the event's spatiotemporal position, in... Figure 3 The event data is plotted as points at the spatiotemporal location of the event (x, y, t).

[0058] By using event data output from EVS camera 11 as pixel values, an event image can be generated using event data within a predetermined frame width starting from the predetermined frame interval for each predetermined frame interval.

[0059] Frame width and frame interval can be specified by time or by the number of event data entries. One of these two parameters can be specified by time, and the other by the number of event data entries.

[0060] Here, when the frame width and frame interval are specified by time and are the same, the frame volumes are in a state of contact with each other without gaps. Furthermore, when the frame interval is greater than the frame width, the frame volumes are arranged with gaps. When the frame width is greater than the frame interval, the frame volumes are arranged in a state of partial overlap.

[0061] For example, an event image can be generated by setting the pixel (pixel value) at the location (x, y) of the event in the frame to white and setting the pixels at other locations in the frame to a predetermined color such as gray.

[0062] Furthermore, when distinguishing the polarity of light intensity change as an event for event data, the generation of frame data can be performed, for example, by setting pixels to white when the polarity is positive, setting pixels to black when the polarity is negative, and setting pixels at other locations in the frame to a predetermined color such as gray.

[0063] Figure 4 This shows an example of an event image capturing a falling spark 24.

[0064] Spark 24 has a brighter light intensity than its surrounding background. Therefore, in a scenario captured by EVS camera 11 where spark 24 falls from the position indicated by the dashed line to the position indicated by the solid line, as... Figure 4 As shown, a brightness change (light intensity change) from dark to bright occurs in the lower region toward which the spark 24 travels, and a positive event occurs. On the other hand, a brightness change (light intensity change) from bright to dark occurs in the upper region opposite to the region toward which the spark 24 travels, and a negative event occurs.

[0065] When generating an image where positively polarized pixels are set to white, negatively polarized pixels are set to black, and pixels at other positions in the frame are set to gray, as a display image to be displayed on monitor 13, the image obtained is... Figure 4 The rightmost image.

[0066] <3. Configuration Examples of Information Processing Devices>

[0067] Figure 5 This is a block diagram showing a detailed configuration example of the information processing device 12.

[0068] It should be noted that, in addition to the EVS camera 11 and monitor 13, in Figure 5 The optional external sensor 14 is also shown.

[0069] The information processing device 12 includes a data acquisition unit 50, an event data processing unit 51, an event data storage unit 52, an image generation unit 53, an image storage unit 54, and an image data processing unit 55. Furthermore, the information processing device 12 includes a grinding stone state estimation unit 56, a camera setting change unit 57, a feature quantity storage unit 58, and an output unit 59.

[0070] The data acquisition unit 50 acquires event data output from the EVS camera 11 at any time and provides the event data to the event data processing unit 51 and the event data storage unit 52.

[0071] The event data processing unit 51 performs predetermined event data processing using the event data provided by the data acquisition unit 50, and provides the processed data to the millstone state estimation unit 56. For example, the event data processing unit 51 calculates the event rate as the frequency of occurrence of event data and provides the event rate to the millstone state estimation unit 56.

[0072] The event data storage unit 52 stores event data provided from the data acquisition unit 50 within a specific time period and provides the event data to the image generation unit 53. The image generation unit 53 uses the event data stored in the event data storage unit 52 to generate event images. Specifically, the image generation unit 53 uses event data within a predetermined frame width starting from a predetermined frame interval from the event data stored in the event data storage unit 52 to generate event images. The event images generated every predetermined frame interval are provided to the image storage unit 54. The image storage unit 54 stores the event images provided from the image generation unit 53.

[0073] The image data processing unit 55 uses the event images stored in the image storage unit 54 to perform predetermined image data processing. For example, the image data processing unit 55 calculates the number of sparks 24, the size of sparks 24, the speed of sparks 24, the flight distance of sparks 24, and the flight angle of sparks 24 within the event images, and provides the calculation results to the grinding stone state estimation unit 56. The number of sparks 24 is, for example, the number of sparks 24 detected within the event images. The size of sparks 24 is, for example, the external size (vertical and horizontal dimensions) of sparks 24 detected within the event images. The speed of sparks 24 is the moving speed calculated from the position of the same spark 24 detected in multiple event images. The flight distance of sparks 24 is the distance from the position where sparks 24 are first detected to the position just before sparks 24 disappear. The flight angle of sparks 24 is the angle between the direction that starts at the position where sparks 24 are first detected and ends at the position where sparks 24 are about to disappear, and the vertically downward direction.

[0074] The information processing device 12 can detect not only the spark 24 but also the coolant liquid 23 as an event, depending on the set value of the event threshold. When the threshold is set so that the coolant liquid 23 is also detected as event data, the image data processing unit 55 also calculates the number of coolant liquid droplets, the size of coolant liquid droplets, and the droplet velocity of coolant liquid 23 based on the event image, and provides the calculation results to the grinding stone state estimation unit 56.

[0075] In the following text, the number of sparks 24, the size of sparks 24, and the speed of sparks 24 may be referred to as the number of sparks, the size of sparks 24, and the speed of sparks 24, and the number of coolant liquid droplets 23, the size of coolant liquid droplets 23, and the speed of coolant liquid droplets 23 may be referred to as the number of droplets, the size of droplets, and the speed of droplets, to distinguish them from each other.

[0076] The millstone state estimation unit 56 estimates the state of the millstone 22 using event processing data provided from the event data processing unit 51 or the image data processing unit 55. Specifically, the millstone state estimation unit 56 determines whether the millstone 22 is clogged by using at least one of the following characteristics: event rate, number of sparks, spark size, and spark speed.

[0077] For example, the millstone state estimation unit 56 determines whether the spark size is equal to or less than a predetermined first state determination threshold VS1, and if it is determined that the spark size is equal to or less than the first state determination threshold VS1, it determines that the millstone 22 is blocked.

[0078] Furthermore, for example, the millstone state estimation unit 56 compares the number of sparks and the spark size with predetermined state determination thresholds. Specifically, the millstone state estimation unit 56 determines whether the number of sparks is equal to or less than a first state determination threshold VS2 and whether the spark size is equal to or less than a second state determination threshold VS3. If it is determined that the number of sparks is equal to or less than the first state determination threshold VS2 and the spark size is equal to or less than the second state determination threshold VS3, the millstone state estimation unit 56 determines that the millstone 22 is blocked.

[0079] If the grinding stone 22 is determined to be clogged, the grinding stone status estimation unit 56 generates an alarm image such as "Clogged and requires maintenance." The alarm image is then output to the display 13 via the output unit 59. Conversely, if the grinding stone 22 is determined to be in a normal state, the grinding stone status estimation unit 56 can generate a display image at a predetermined frame rate and output the display image to the display 13 via the output unit 59.

[0080] Furthermore, the grinding stone state estimation unit 56 also has the function of adjusting the event threshold of the EVS camera 11 based on event processing data provided from the event data processing unit 51 or the image data processing unit 55. For example, the grinding stone state estimation unit 56 increases or decreases the event threshold based on the event rate provided from the event data processing unit 51 to instruct the camera setting change unit 57 to do so. The camera setting change unit 57 changes the event threshold of the EVS camera 11 based on the instruction from the grinding stone state estimation unit 56 to increase or decrease the event threshold.

[0081] The feature quantity storage unit 58 is a storage unit that stores the feature quantities obtained by the grinding stone state estimation unit 56 from the event data processing unit 51 or the image data processing unit 55.

[0082] Output unit 59 outputs the alarm image provided by grinding stone state estimation unit 56 to display 13. In addition, output unit 59 can also output event images and display images to display 13.

[0083] The information processing device 12 is configured as described above and is capable of estimating the state of the grinding stone 22 based on event data output from the EVS camera 11, and detecting, for example, the occurrence of blockage in the grinding stone 22. The information processing device 12 can prompt the operator to perform maintenance by displaying an alarm image on the display 13.

[0084] Furthermore, the information processing device 12 can be connected to the external sensor 14, and also estimates the state of the grinding stone 22 by using event data output from the EVS camera 11 and sensor data obtained from the external sensor 14. The external sensor 14 can be, for example, a microphone that detects sound during grinding, a far-infrared sensor that measures temperature, etc. Needless to say, the external sensor 14 can be a sensor other than a microphone or a far-infrared sensor.

[0085] When the external sensor 14 is connected to the information processing device 12, the sensor data generated by the external sensor 14 is provided to the grinding stone state estimation unit 56. The grinding stone state estimation unit 56 estimates the state of the grinding stone 22 by using the sensor data provided from the external sensor 14 and the event processing data provided from the event data processing unit 51 or the image data processing unit 55.

[0086] <4. Relationship between measurement parameters and physical quantities>

[0087] Figure 6 It is a diagram showing the relationship between parameters (measurement parameters) that can be measured by the EVS camera 11 (event sensor) and physical quantities related to the grinding process of the machine tool 21.

[0088] exist Figure 6 In the image, the items that the EVS camera 11 can measure are surrounded by thick lines.

[0089] Examples of measuring devices used to determine the necessity of maintenance for machine tool 21 include Figure 6 The rightmost column describes the surface roughness meter, RGB camera, thermocouple, and thermal imaging. In this embodiment, the EVS camera 11 (event sensor) is used instead of these measuring devices.

[0090] The EVS camera 11 can generate and output event data. The event data includes event data for the spark 24 and event data for the coolant liquid 23. Sometimes events caused by ambient light or device vibration are detected. Because events caused by ambient light or device vibration correspond to noise, such events can be excluded by appropriately setting the event threshold.

[0091] In the event data of spark 24, the number of sparks, spark size, spark velocity, and spark breakage mode can be measured as parameters. The spark breakage mode is a classification indicating the breakage (breakage mode) characteristics of spark 24. The spark breakage mode varies depending on the material of the workpiece W. The material of workpiece W can be specified by detecting the spark breakage mode.

[0092] The number of sparks is related to the abrasive grain shedding frequency, the grinding peripheral speed, and the feed rate as machining conditions. The spark size is related to the abrasive grain size of the grinding stone 22, as well as the feed rate and cutting depth as machining conditions. The spark speed is related to the grinding peripheral speed. The abrasive grain shedding frequency is related to the adhesion of the binder in the grinding stone 22 and the porosity of the pores. The grinding peripheral speed is related to the peripheral speed of the workpiece W and the peripheral speed of the grinding stone 22 as machining conditions.

[0093] In the event data of coolant liquid 23, the number of droplets, droplet size, and droplet velocity can be measured as measurement parameters. The number of droplets, droplet size, and droplet velocity are related to the flow rate of coolant liquid 23.

[0094] The maintenance of the grinding stone 22 of machine tool 21, especially the clogging of the grinding stone 22, is highly correlated with the spark size, which can be measured by event data from spark 24. The spark size is highly correlated with the particle size of the abrasive particles in terms of physical quantities.

[0095] Figure 7 This shows the measurable measurement parameters based on event data and in... Figure 6 The table shows the correlation between physical quantities and measured parameters, represented by thick lines.

[0096] exist Figure 7 In the table, a positive correlation between the physical quantity value shown on the left and the measurement parameter shown on the top is indicated by "+", a negative correlation by "-", and a correlation other than positive or negative is indicated by a circle.

[0097] For example, the abrasive grain size and spark size of grinding stone 22 are related as follows: as the abrasive grain size increases, the spark size increases. The adhesion of the binder and the number of sparks are related as follows: the number of sparks increases with increasing adhesion. The porosity of the pores is related to the number and size of sparks as follows: as the porosity increases, both the number and size of sparks decrease.

[0098] For example, the flow rate of coolant liquid 23 is related to the number of droplets and the droplet velocity in such a way that the number of droplets and the droplet velocity also increase with the increase of flow rate.

[0099] According to such Figure 7 The correlation between measurable parameters and physical quantities shown in the event data allows the grinding stone condition estimation unit 56 to estimate physical quantities and determine maintenance time based on the data processing results of the event data.

[0100] <5. Flowchart for determining maintenance time>

[0101] Next, we will refer to Figure 8The flowchart describes the maintenance time determination process performed by the information processing system 1. For example, this process begins when the EVS camera 11 and the information processing device 12 are activated (powered on).

[0102] First, in step S11, the data acquisition unit 50 acquires event data output from the EVS camera 11 at any time and provides the event data processing unit 51 and the event data storage unit 52.

[0103] In step S12, the event data processing unit 51 performs predetermined event data processing using the event data provided by the data acquisition unit 50, and provides the processed data to the millstone state estimation unit 56. For example, the event data processing unit 51 calculates the event rate as the frequency of occurrence of event data and provides the event rate to the millstone state estimation unit 56.

[0104] In step S13, the event data storage unit 52 stores event data provided by the data acquisition unit 50 within a specific time period and provides the event data to the image generation unit 53. The image generation unit 53 generates an event image using the event data stored in the event data storage unit 52 and provides the event image to the image storage unit 54.

[0105] In step S14, the image data processing unit 55 performs predetermined image data processing using the event image stored in the image storage unit 54. For example, the image data processing unit 55 calculates the number of sparks 24, the size of sparks 24, the speed of sparks 24, the flight distance of sparks 24, and the flight angle of sparks 24 in the event image, and provides the calculation results to the grinding stone state estimation unit 56.

[0106] In step S15, the millstone state estimation unit 56 performs millstone state estimation processing to estimate the state of the millstone 22 using event processing data provided from the event data processing unit 51 or the image data processing unit 55. For example, as millstone state estimation processing, the millstone state estimation unit 56 determines whether the spark size is equal to or less than a first state determination threshold VS1. Optionally, as millstone state estimation processing, the millstone state estimation unit 56 determines whether the number of sparks is equal to or less than a first state determination threshold VS2 and whether the spark size is equal to or less than a second state determination threshold VS3.

[0107] In step S16, the grinding stone state estimation unit 56 determines whether the grinding stone 22 is blocked based on the result of the grinding stone state estimation processing.

[0108] If it is determined in step S16 that the grinding stone 22 is not clogged, the process returns to step S11, and the processes in steps S11 to S16 are executed again. It should be noted that when the grinding stone 22 is in a normal state, that is, when it is not clogged, the display image generated based on the event data, the event image generated by the image generation unit 53, etc., can be provided to the display 13 and displayed through the output unit 59.

[0109] On the other hand, if it is determined in step S16 that the grinding stone 22 is clogged, the process proceeds to step S17, and the grinding stone status estimation unit 56 provides an alarm for the clogged grinding stone 22. For example, the grinding stone status estimation unit 56 generates an alarm image such as "A clog has occurred and maintenance is required." The alarm image is then output to the display 13 via the output unit 59. The display 13 displays the alarm image provided by the information processing device 12.

[0110] As described above, the maintenance time determination process of the information processing system 1 is performed. Through the above process, the operator confirms that the alarm image displayed on the display 13 has arrived and performs, for example, the grinding stone 22.

[0111] <Grinding Stone State Estimation Process Using a Learned Model>

[0112] In the above-described millstone state estimation process, at least one of the following is used as a characteristic quantity: the number of sparks 24, the size of sparks 24, the speed of sparks 24, the flight distance of sparks 24, the flight angle of sparks 24, the number of droplets of coolant liquid 23, the droplet size of coolant liquid 23, or the droplet speed of coolant liquid 23. The state of millstone 22 is estimated by a threshold determination process that compares the characteristic quantity with a predetermined threshold.

[0113] Optionally, as a grinding stone state estimation process for estimating the state of the grinding stone 22, the state of the grinding stone 22 can be estimated, and the maintenance time can be determined by using a learning model generated through machine learning. For example, the grinding stone state estimation unit 56 generates a learning model by using event data obtained during grinding with, for example, a clogged and maintenance-required grinding stone 22 and event data obtained during grinding with a grinding stone 22 in a normal state (maintenance-free state), using the necessity of maintenance as training data. The grinding stone state estimation unit 56 estimates the necessity of maintaining the grinding stone 22 based on the input event data using the generated learning model. Alternatively, instead of the event data itself, features such as the number of sparks 24, the size of sparks 24, the speed of sparks 24, the flight distance of sparks 24, and the flight angle of sparks 24 can be used as training data for generating the learning model. The learning model can be trained to not only determine the necessity of maintenance but also the state of the grinding stone 22, such as clogs, dulling, or detachment.

[0114] <Grinding stone condition estimation processing using external sensor data>

[0115] Furthermore, when the external sensor 14 is connected to the information processing device 12, in addition to the data processing results of the event data, the state of the grinding stone 22 can also be estimated using sensor data obtained by the external sensor 14. The grinding stone state estimation process using the data processing results of the event data and the sensor data can be a threshold determination process or a determination process using a learning model.

[0116] <6. Flowchart of Threshold Update Processing>

[0117] Next, we will refer to Figure 9 The flowchart describes the threshold update process used to dynamically change the event threshold. For example, this process is compared with a reference... Figure 8 The described maintenance time determination process starts together with and is executed in parallel with the maintenance time determination process.

[0118] First, in step S31, the millstone state estimation unit 56 acquires the data processing results of the event data or event image. The processing in step S31 includes parallel execution... Figure 8 The maintenance time determination process is in progress and can therefore be largely eliminated. In addition, the millstone state estimation unit 56 can also acquire the event data itself output from the EVS camera 11 via the event data processing unit 51.

[0119] In step S32, the grinding stone state estimation unit 56 calculates the degree of influence of the coolant liquid 23 using the acquired data processing results. For example, when using the event rate provided by the event data processing unit 51, the grinding stone state estimation unit 56 can calculate the degree of influence of the coolant liquid 23 based on the event rate in the state where no spark 24 is emitted. Furthermore, for example, when using the data processing results of the event image, the grinding stone state estimation unit 56 can calculate the degree of influence of the coolant liquid 23 based on the ratio of the number of sparks to the number of droplets. The spark 24 and the coolant liquid 23 can be distinguished, for example, by size.

[0120] In step S33, the grinding stone state estimation unit 56 determines whether to change the event threshold. For example, if it is desired to detect only spark 24 from the current state where both spark 24 and coolant liquid 23 are detected as events, the grinding stone state estimation unit 56 determines to change the event threshold. In this case, the event threshold is adjusted to a value greater than the current value. Alternatively, if it is desired to detect both spark 24 and coolant liquid 23 from the current state where only spark 24 is detected, the event threshold is determined to change. In this case, the event threshold is adjusted to a value less than the current value.

[0121] If it is determined in step S33 that the event threshold has not changed, the process returns to step S31, and the above steps S31 to S33 are executed again.

[0122] On the other hand, if it is determined in step S33 that the event threshold has changed, the process proceeds to step S34, and the grinding stone state estimation unit 56 instructs the camera setting change unit 57 to increase or decrease the event threshold. The camera setting change unit 57 sets a new event threshold by providing the new event threshold to the EVS camera 11. The new event threshold is, for example, a value obtained by changing the event threshold by a predetermined change width in the indicated increase or decrease direction.

[0123] Based on the threshold update process described above, the event threshold can be adjusted in parallel with the grinding stone state estimation process based on the event detection state. In the information processing device 12, for example, by setting an operating mode, it can be pre-specified that both spark 24 and coolant liquid 23 are detected as events, or only spark 24 is detected as an event.

[0124] <7. Second Embodiment of the Information Processing System>

[0125] Next, a second embodiment applying this technology will be described.

[0126] exist Figure 1In the first embodiment of the information processing system shown, the EVS camera 11 detects events such as changes in the brightness of the spark 24 and outputs event data to the information processing device 12, and the information processing device 12 performs a process of estimating the state of the grinding wheel 22 by using the event data.

[0127] On the other hand, in the information processing system 1 of the second embodiment described below, a process of estimating the state of the grinding wheel 22 by using event data is also performed in the EVS camera. In other words, the EVS camera 11 and the information processing device 12 in the first embodiment are replaced by Figure 10 one EVS camera 300 shown.

[0128] <Configuration Example of EVS Camera>

[0129] In Figure 10 the EVS camera 300 shown is an imaging device including an event sensor and a processing unit that performs the functions of the information processing device 12 of the first embodiment. The EVS camera 300 is installed at the same position as the Figure 1 EVS camera 11 shown, detects changes in the brightness of the spark 24 or the coolant liquid 23 as events, and generates event data. In addition, the EVS camera 300 performs a grinding wheel state estimation process of estimating the state of the grinding wheel 22 based on the event data, and outputs a maintenance alarm based on the result of the grinding wheel state estimation process. For example, in the case where maintenance is determined to be necessary, the EVS camera 300 causes the display 13 to display an alarm image such as "Clogging has occurred, maintenance is required". In addition, the EVS camera 300 can generate a display image to be monitored by the operator based on the event data, and cause the display 13 to display the display image.

[0130] The EVS camera 300 includes an optical unit 311, an imaging element 312, a control unit 313, and a data processing unit 314.

[0131] The optical unit 311 collects light from an object and causes the light to enter the imaging element 312. The imaging element 312 performs photoelectric conversion on the incident light incident via the optical unit 311 to generate event data, and provides the event data to the data processing unit 314. The imaging element 312 is a light receiving element that outputs event data indicating the occurrence of an event having a brightness change in a pixel as an event.

[0132] The control unit 313 controls the imaging element 312. For example, the control unit 313 instructs the imaging element 312 to start and end imaging.

[0133] The data processing unit 314 includes, for example, a field-programmable gate array (FPGA), a digital signal processor (DSP), a microprocessor, etc., and performs the processing performed by the information processing device 12 in the first embodiment. The data processing unit 314 includes an event data processing unit 321 and a recording unit 322. For example, the event data processing unit 321 performs event data processing using event data provided from the imaging element 312, image data processing using event images, and grinding stone state estimation processing to estimate the state of the grinding stone 22. The recording unit 322 corresponds to the event data storage unit 52, image storage unit 54, and feature storage unit 58 in the first embodiment, and records and accumulates predetermined data in a predetermined recording medium as needed.

[0134] <Imaging element configuration example>

[0135] Figure 11 This is a block diagram illustrating a schematic configuration example of the imaging element 312.

[0136] The imaging element 312 includes a pixel array unit 341, a driving unit 342, a Y arbitrator 343, an X arbitrator 344, and an output unit 345.

[0137] In pixel array unit 341, multiple pixels 361 are arranged in a two-dimensional grid. Each pixel 361 includes a photodiode 371 as a photoelectric conversion element and an address event detection circuit 372. When a change exceeding a predetermined threshold occurs in the photocurrent, which is an electrical signal generated by photoelectric conversion through photodiode 371, the address event detection circuit 372 detects the change in photocurrent as an event. Upon detecting an event, the address event detection circuit 372 outputs a request to output event data indicating the occurrence of the event to Y arbitrator 343 and X arbitrator 344.

[0138] The driving unit 342 drives the pixel array unit 341 by providing control signals to each pixel 361 of the pixel array unit 341.

[0139] Y arbitrator 343 arbitrates a request from pixel 361 in the same row of pixel array unit 341 and returns a permission or disallowment response indicating the output of event data to pixel 361 that sent the request. X arbitrator 344 arbitrates a request from pixel 361 in the same column of pixel array unit 341 and returns a permission or disallowment response indicating the output of event data to pixel 361 that sent the request. Pixel 361 that has received permission responses from both Y arbitrator 343 and X arbitrator 344 can output event data to output unit 345.

[0140] Note that the imaging element 312 may include only one of the Y arbitrator 343 and the X arbitrator 344. For example, in the case of including only the X arbitrator 344, data of all pixels 361 in the same column, including the pixel 361 that sent the request, is transmitted to the output unit 345. Then, in subsequent stages, the output unit 345 or the data processing unit 314 ( Figure 10 In the process, only event data for pixels 361 whose events have actually occurred is selected. With only the Y arbitrator 343 included, pixel data is transmitted to the output unit 345 row by row, and in subsequent stages, only event data for necessary pixels 361 is selected.

[0141] The output unit 345 performs necessary processing on the event data output from each pixel 361 constituting the pixel array unit 341, and provides the processed event data to the data processing unit 314. Figure 10 ).

[0142] <Configuration Example of Address Event Detection Circuit>

[0143] Figure 12 This is a block diagram showing a configuration example of the address event detection circuit 372.

[0144] Address event detection circuit 372 includes current-to-voltage conversion circuit 381, buffer 382, ​​subtractor 383, quantizer 384, and transmission circuit 385.

[0145] The current-to-voltage conversion circuit 381 converts the photocurrent from the corresponding photodiode 371 into a voltage signal. The current-to-voltage conversion circuit 381 generates a voltage signal corresponding to the logarithm of the photocurrent and outputs the voltage signal to the buffer 382.

[0146] Buffer 382 buffers the voltage signal from the current-to-voltage conversion circuit 381 and outputs the voltage signal to the subtractor 383. Buffer 382 ensures isolation from noise generated by switching operations in subsequent stages and increases the driving force for driving subsequent stages. Note that buffer 382 can be omitted.

[0147] Subtractor 383 reduces the level of the voltage signal from buffer 382 according to the control signal from drive unit 342. Subtractor 383 outputs the reduced voltage signal to quantizer 384.

[0148] Quantizer 384 quantizes the voltage signal from subtractor 383 into a digital signal and provides this digital signal as event data to transmission circuit 385. Transmission circuit 385 transmits (outputs) the event data to output unit 345. That is, transmission circuit 385 provides a request to Y arbitrator 343 and X arbitrator 344 to output event data. Then, in response to receiving a response from Y arbitrator 343 and X arbitrator 344 indicating that output of event data is permitted, transmission circuit 385 transmits the event data to output unit 345.

[0149] <Detailed Configuration Example of Address Event Detection Circuit>

[0150] Figure 13 This is a circuit diagram showing the detailed configuration of the current-to-voltage conversion circuit 381, the subtractor 383, and the quantizer 384. Figure 13 The image also shows a photodiode 371 connected to a current-to-voltage conversion circuit 381.

[0151] The current-to-voltage conversion circuit 381 includes FETs 411 to 413. For example, N-type metal-oxide-semiconductor (NMOS) FETs can be used as FETs 411 and FET 413, and for example, P-type metal-oxide-semiconductor (PMOS) FETs can be used as FET 412.

[0152] Photodiode 371 receives incident light, performs photoelectric conversion, and generates and allows the flow of photocurrent as an electrical signal. Current-to-voltage conversion circuit 381 converts the photocurrent from photodiode 371 into a voltage (hereinafter also referred to as photovoltage) VLOG corresponding to the logarithm of the photocurrent, and outputs this voltage VLOG to buffer 382.

[0153] The source of FET 411 is connected to the gate of FET 413, and the photocurrent from photodiode 371 flows through the connection point between the source of FET 411 and the gate of FET 413. The drain of FET 411 is connected to the power supply VDD, and its gate is connected to the drain of FET 413.

[0154] The source of FET 412 is connected to the power supply VDD, and its drain is connected to the junction between the gate of FET 411 and the drain of FET 413. A predetermined bias voltage Vbias is applied to the gate of FET 412. The source of FET 413 is grounded.

[0155] The drain of FET 411 is connected to the power supply VDD side and is a source follower. Photodiode 371 is connected to the source of FET 411, which is also a source follower. This connection allows photocurrent to flow through the drain to the source of FET 411 via the charge generated by the photoelectric conversion of photodiode 371. FET 411 operates in the subthreshold region, and a photovoltage VLOG corresponding to the logarithm of the photocurrent flowing through FET 411 appears at the gate of FET 411. As described above, in photodiode 371, the photocurrent from photodiode 371 is converted into a photovoltage VLOG corresponding to the logarithm of the photocurrent through FET 411.

[0156] The photovoltage VLOG is output from the junction between the gate of FET 411 and the drain of FET 413 via buffer 382 to subtractor 383.

[0157] For the photovoltage VLOG from the current-to-voltage conversion circuit 381, the subtractor 383 calculates the difference between the photovoltage at the current time and the photovoltage at a timing point that differs from the current time by a small amount of time, and outputs a differential signal Vdiff corresponding to the difference.

[0158] The subtractor 383 includes a capacitor 431, an operational amplifier 432, a capacitor 433, and a switch 434. The quantizer 384 includes comparators 451 and 452.

[0159] One end of capacitor 431 is connected to the output of buffer 382, ​​and the other end is connected to the input of operational amplifier 432. Therefore, the photovoltage VLOG is input to the (inverting) input of operational amplifier 432 via capacitor 431.

[0160] The output of operational amplifier 432 is connected to the non-inverting input (+) of comparators 451 and 452 of quantizer 384.

[0161] One end of capacitor 433 is connected to the input terminal of operational amplifier 432, and the other end is connected to the output terminal of operational amplifier 432.

[0162] Switch 434 is connected to capacitor 433 to turn the connection between the two ends of capacitor 433 on / off. Switch 434 turns the connection between the two ends of capacitor 433 on / off according to the control signal of drive unit 342.

[0163] Capacitor 433 and switch 434 constitute a switched capacitor. When switch 434, which has been open, is temporarily turned on and then turned off again, capacitor 433 is reset to a state where the charge has been discharged and the charge can be re-accumulated.

[0164] When switch 434 is turned on, the photovoltage VLOG of capacitor 431 on the side of photodiode 371 is represented by Vinit, and the capacitance (static capacitance) of capacitor 431 is represented by C1. The input terminal of operational amplifier 432 is virtually grounded, and when switch 434 is turned on, the charge Qinit accumulated in capacitor 431 is represented by formula (1).

[0165] Qinit = C1 × Vinit (1)

[0166] Furthermore, when switch 434 is turned on, the two ends of capacitor 433 are short-circuited, causing the charge accumulated in capacitor 433 to become 0.

[0167] Subsequently, when the photovoltage VLOG of capacitor 431 on the photodiode 371 side is expressed as Vafter when switch 434 is open, the charge Qafter accumulated in capacitor 431 when switch 434 is open is expressed by formula (2).

[0168] Qafter=C1×Vafter (2)

[0169] When the capacitance of capacitor 433 is represented as C2, the charge Q2 accumulated in capacitor 433 is represented by formula (3) by using the differential signal Vdiff, which is the output voltage of operational amplifier 432.

[0170] Q2=-C2×Vdiff (3)

[0171] Before and after switch 434 is turned off, the total charge of capacitor 431 and capacitor 433 remains unchanged, thus establishing formula (4).

[0172] Qinit=Qafter+Q2 (4)

[0173] When formulas (1) to (3) are substituted into formula (4), formula (5) is obtained.

[0174] Vdiff=-(C1 / C2)×(Vafter-Vinit) (5)

[0175] According to formula (5), subtractor 383 subtracts the photovoltages Vafter and Vinit, thus calculating the differential signal Vdiff corresponding to the difference between the photovoltages Vafter and Vinit (Vafter-Vinit). According to formula (5), the subtraction gain of subtractor 383 is C1 / C2. Therefore, subtractor 383 uses the voltage output obtained by multiplying the change in photovoltage VLOG after resetting capacitor 433 by C1 / C2 as the differential signal Vdiff.

[0176] Subtractor 383 outputs differential signal Vdiff by turning switch 434 on and off using a control signal output from drive unit 342.

[0177] The differential signal Vdiff output from subtractor 383 is provided to the non-inverting inputs (+) of comparators 451 and 452 of quantizer 384.

[0178] Comparator 451 compares the differential signal Vdiff from subtractor 383 with the positive threshold Vrefp input to the inverting input (-). Comparator 451 outputs the quantized value of the differential signal Vdiff to transmission circuit 385 as a detection signal DET(+), indicating whether the differential signal Vdiff has exceeded the high (H) or low (L) level of the positive threshold Vrefp.

[0179] Comparator 452 compares the differential signal Vdiff from subtractor 383 with the negative-side threshold Vrefn input to the inverting input (-). Comparator 452 outputs the quantized value of the differential signal Vdiff to transmission circuit 385 as a detection signal DET(-), indicating whether the differential signal Vdiff has exceeded the high (H) or low (L) level of the negative-side threshold Vrefn.

[0180] Figure 14 It shows Figure 13 A more detailed circuit configuration example of the current-to-voltage conversion circuit 381, buffer 382, ​​subtractor 383, and quantizer 384 shown is provided.

[0181] Figure 15 This is a circuit diagram showing another configuration example of the quantizer 384.

[0182] exist Figure 14 The quantizer 384 shown compares the differential signal Vdiff from the subtractor 383 with the positive-side threshold (voltage) Vrefp and the negative-side threshold (voltage) Vrefn constant, and outputs the comparison result.

[0183] on the other hand, Figure 15 The quantizer 384 includes a comparator 453 and a switch 454, and outputs a comparison result that is compared with either of two threshold (voltage) VthON and VthOFF switched by the switch 454.

[0184] Switch 454 is connected to the inverting input (-) of comparator 453 and selects terminal a or b according to the control signal from drive unit 342. A threshold voltage VthON is provided to terminal a, and a threshold voltage VthOFF (<VthON) is provided to terminal b. Therefore, either voltage VthON or VthOFF is provided to the inverting input of comparator 453.

[0185] Comparator 453 compares the differential signal Vdiff from subtractor 383 with voltage VthON or VthOFF, and outputs the detection signal DET, which indicates the comparison result (H level or L level), as the quantized value of the differential signal Vdiff to transmission circuit 385.

[0186] Figure 16 Showing the adoption Figure 15 A more detailed circuit configuration example of the current-to-voltage conversion circuit 381, buffer 382, ​​subtractor 383, and quantizer 384 is shown in the figure.

[0187] exist Figure 16 In the circuit configuration, in addition to voltages VthON and VthOFF, a terminal VAZ (AutoZero) for initialization is added as a terminal of switch 454. When the H (high) level initialization signal AZ is provided to the gate of FET 471, which serves as an N-type MOS (NMOS) FET in subtractor 383, switch 454 of quantizer 384 selects terminal VAZ and performs the initialization operation. Subsequently, switch 454 selects either the terminal for voltage VthON or the terminal for voltage VthOFF based on the control signal from drive unit 342, and outputs a detection signal DET indicating the comparison result with the selected threshold from quantizer 384 to transmission circuit 385.

[0188] The maintenance time determination process and threshold update process in the second embodiment are similar to those in the first embodiment, except that the maintenance time determination process and threshold update process are performed by the EVS camera 300 itself, not by the information processing device 12. Therefore, the sparks 24 and coolant liquid 23 generated during grinding can be detected as events, and the maintenance time can be accurately determined.

[0189] <8. Conclusion>

[0190] According to the embodiment of the information processing system 1 described above, maintenance time can be more easily determined by using an event sensor (EVS camera 11 or EVS camera 300), which detects changes in brightness such as sparks 24 as events and outputs the events asynchronously. Furthermore, the event threshold can be dynamically changed based on the event detection status.

[0191] Although the above embodiments have described an example in which machine tool 21 is a grinding machine, machine tool 21 can be any machine that performs processes such as cutting, grinding, slicing, forging, or bending.

[0192] <9. Computer Configuration Examples>

[0193] The series of processes performed by the aforementioned information processing device 12 can be executed by hardware or software. In the case where the series of processes are executed by software, programs constituting the software are installed in the computer. Examples of computers include, for instance, microcomputers built into dedicated hardware, and general-purpose personal computers capable of performing various functions by installing various programs.

[0194] Figure 17 This is a block diagram illustrating a hardware configuration example of a computer, which is an information processing device that performs the above-described series of processes through a program.

[0195] In a computer, the central processing unit (CPU) 501, read-only memory (ROM) 502, and random access memory (RAM) 503 are interconnected via bus 504.

[0196] The input / output interface 505 is further connected to the bus 504. The input unit 506, output unit 507, storage unit 508, communication unit 509, and driver 510 are connected to the input / output interface 505.

[0197] Input unit 506 includes a keyboard, mouse, microphone, touch panel, input terminal, etc. Output unit 507 includes a display, speaker, output terminal, etc. Storage unit 508 includes a hard disk, RAM disk, non-volatile memory, etc. Communication unit 509 includes a network interface, etc. Driver 510 drives removable recording media 511 such as disk, optical disk, magneto-optical disk, or semiconductor memory.

[0198] In the computer configured as described above, for example, the CPU 501 loads a program stored in the storage unit 508 into the RAM 503 via the input / output interface 505 and the bus 504 and executes the program, thereby performing the series of processes described above. The RAM 503 also appropriately stores the data required by the CPU 501 to perform various processes.

[0199] For example, a program executed by a computer (CPU 501) can be provided by recording it on a removable recording medium 511, which serves as the encapsulation medium. Alternatively, the program can be provided via wired or wireless transmission media such as a local area network, the Internet, or digital broadcasting.

[0200] In a computer, a program can be installed in the storage unit 508 via the input / output interface 505 by attaching the removable recording medium 511 to the drive 510. Alternatively, the program can be received by the communication unit 509 via a wired or wireless transmission medium and installed in the storage unit 508. Additionally, the program can be pre-installed in the ROM 502 or the storage unit 508.

[0201] It should be noted that a program executed by a computer may be a program for execution in a time sequence as described in this specification, or a program for parallel processing or necessary timing processing, such as when a call is made.

[0202] The embodiments of this technology are not limited to the above embodiments, and various modifications can be made without departing from the spirit of this technology.

[0203] For example, it is possible to appropriately adopt all or part of the combinations of the various embodiments described above.

[0204] Furthermore, each step described in the flowchart above can be performed by a single device or by multiple devices in a shared manner.

[0205] Furthermore, when multiple processes are included in a single step, the multiple processes included in a single step can be executed by a single device or by multiple devices in a shared manner.

[0206] It should be noted that the effects described in this specification are merely illustrative and not limiting, and effects other than those described in this specification may be provided.

[0207] It should be noted that this technology can have the following configurations.

[0208] (1) An information processing device, comprising:

[0209] The state estimation unit estimates the state of the grinding stone by using event data provided by the event sensor and outputs the estimation result. The event sensor outputs the time change of the electrical signal obtained by photoelectric conversion of light signal as event data.

[0210] (2) The information processing device according to (1),

[0211] The state estimation unit estimates the state of the grinding stone by using the event data captured from the sparks generated between the grinding stone and the workpiece, and outputs the estimation result.

[0212] (3) The information processing device according to (1) or (2),

[0213] The state estimation unit estimates the state of the grinding stone based on the feature quantities of the event data and outputs the estimation result.

[0214] (4) The information processing device according to (3),

[0215] Among them, the characteristic quantity of event data is the event rate.

[0216] (5) The information processing apparatus according to (3) further includes:

[0217] The image generation unit generates an event image from the event data.

[0218] The feature quantity of the event data is the feature quantity detected from the event image.

[0219] (6) The information processing device according to (5),

[0220] The event data features include at least one of the following: the number of sparks, the size of the sparks, the speed of the sparks, the flight distance of the sparks, or the flight angle of the sparks.

[0221] (7) The information processing apparatus according to (5) or (6),

[0222] The event data features include at least one of the following: the number of coolant liquid droplets, the size of the coolant liquid droplets, or the velocity of the coolant liquid droplets.

[0223] (8) The information processing apparatus according to any one of (1) to (7),

[0224] The state estimation unit outputs an alarm based on the estimation result.

[0225] (9) The information processing apparatus according to any one of (1) to (8),

[0226] The state estimation unit adjusts the event threshold based on event data in parallel with the processing of estimating the state of the grinding stone.

[0227] (10) The information processing apparatus according to any one of (1) to (9),

[0228] The state estimation unit estimates the state of the grinding stone by using a learning model generated by the machine learning using the event data, and outputs the estimated result.

[0229] (11) The information processing apparatus according to any one of (1) to (10),

[0230] The state estimation unit estimates the state of the grinding stone by using sensor data acquired from external sensors and the event data, and outputs the estimation result.

[0231] Reference Symbol List

[0232] 1. Information processing system; 11. EVS camera; 12. Information processing device; 13. Display.

[0233] 14 External Sensors 21 Machine Tools 22 Grinding Stones 23 Coolant Liquid 24 Spark 50 Data Acquisition Unit 51 Event Data Processing Unit 52 Event Data Storage Unit

[0234] 53 Image generation unit; 54 Image storage unit; 55 Image data processing unit

[0235] 56 Grinding stone state estimation unit; 57 Camera setting change unit; 58 Feature quantity storage unit; 59 Output unit; 300 EVS camera; 311 Optical unit; 312 Imaging element.

[0236] 313 Control Unit; 314 Data Processing Unit; 321 Event Data Processing Unit

[0237] 322 Recording Unit; 501 CPU; 502 ROM; 503 RAM; 508 Storage Unit

Claims

1. An information processing apparatus, comprising: The state estimation unit estimates the state of the grinding stone by using event data provided from an event sensor and outputs the estimation result. The event sensor detects sparks generated at the contact portion between the grinding stone and the workpiece during grinding and changes in light caused by coolant liquid dripping onto the contact portion between the grinding stone and the workpiece, and outputs the time change of an electrical signal obtained by photoelectric conversion as the event data.

2. The information processing device according to claim 1, in, The state estimation unit estimates the state of the grinding stone by using the event data captured from the sparks generated between the grinding stone and the workpiece, and outputs the estimation result.

3. The information processing device according to claim 1, in, The state estimation unit estimates the state of the grinding stone based on the feature quantities of the event data and outputs the estimation result.

4. The information processing device according to claim 3, in, The characteristic quantity of the event data is the event rate.

5. The information processing apparatus according to claim 3, further comprising: The image generation unit generates an event image from the event data. The feature quantity of the event data is the feature quantity detected from the event image.

6. The information processing apparatus according to claim 5, in, The features of the event data include at least one of the following: the number of sparks, the size of the sparks, the speed of the sparks, the flight distance of the sparks, and the flight angle of the sparks.

7. The information processing apparatus according to claim 5, in, The characteristic quantities of the event data include at least one of the following: the number of coolant liquid droplets, the size of the coolant liquid droplets, and the velocity of the coolant liquid droplets.

8. The information processing apparatus according to claim 1, in, The state estimation unit outputs an alarm based on the estimation result.

9. The information processing apparatus according to claim 1, in, The state estimation unit adjusts the event threshold based on the event data in parallel with the processing that estimates the state of the grinding stone.

10. The information processing apparatus according to claim 1, in, The state estimation unit estimates the state of the grinding stone by using a learning model generated by machine learning using the event data, and outputs the estimated result.

11. The information processing apparatus according to claim 1, in, The state estimation unit estimates the state of the grinding stone by using sensor data acquired from external sensors and the event data, and outputs the estimation result.

Images