WO2019059343A1

WO2019059343A1 - Workpiece information processing device and recognition method of workpiece

Info

Publication number: WO2019059343A1
Application number: PCT/JP2018/035021
Authority: WO
Inventors: 博明大庭
Original assignee: Ntn株式会社
Priority date: 2017-09-22
Filing date: 2018-09-21
Publication date: 2019-03-28
Also published as: JP2019057250A

Abstract

This workpiece information processing device (1) used to recognize the shape of a workpiece is provided with: a learning data creation unit (3) which creates training data necessary for learning on the basis of design information on a workpiece (8); a workpiece learning unit (5) which learns the training data; and a data conservation unit (7) which conserves the training data and the learnt result in the workpiece learning unit (5). Preferably, the workpiece information processing device (1) is further provided with a workpiece data creation unit (2) which creates, from the design information, workpiece data including a basic image and coordinate data on the workpiece. The learning data creation unit (3) creates the training data by changing the basic image. Accordingly, the workpiece information processing device can be implemented by mechanical learning without requiring trial and error of a human or specialized knowledge in the selection of an image feature or an optimal condition necessary to recognize the shape of the workpiece.

Description

Work information processing apparatus and work recognition method

The present invention relates to a workpiece information processing apparatus that recognizes the shape of a workpiece that is an object of machining or assembly.

In machining or assembly operations, workpieces that are objects are often picked up automatically by a robot or an assembly apparatus and the like and set in a housing or the like of a processing apparatus or assembly. At the time of such pickup, it is necessary to control the pickup arm by recognizing the shape and posture of the work. For example, a workpiece information processing apparatus is used for workpiece posture detection, shape recognition, and the like which are necessary in the work of taking out a workpiece by a robot, an assembly apparatus or the like.

JP-A-10-332333 (patent document 1) and JP-A-11-066321 (patent document 2) photograph a work stacked one upon another with a camera, differentiate the obtained image and extract the outline A work information processing apparatus to perform is disclosed.

In the apparatus disclosed in Japanese Patent Application Laid-Open No. 10-332333 (Patent Document 1), a curvature is obtained from the extracted contour line, and a pixel having a locally maximum curvature is detected as a vertex. A master pattern is an image captured under ideal conditions in advance, and an image of a workpiece actually captured by a camera is used as an input pattern. Master patterns with equal curvatures are associated with vertices of the input pattern, and the number of associated vertices is the largest The amount of deviation of the planar position and the rotation angle between the contour of the input pattern and the contour of the master pattern is detected.

In the apparatus disclosed in Japanese Patent Application Laid-Open No. 11-066321 (Patent Document 2), an image captured in advance under ideal conditions is used as a template, and an image of a work actually captured by a camera is used as an input image. The tangent angle of each pixel on the contour line is determined, the template and the histogram of the tangent angle of the input image are superimposed, and the deviation of the angle when they are most coincident is detected as the rotation angle.

Japanese Patent Application Laid-Open No. 10-332333 Japanese Patent Application Laid-Open No. 11-066321

The apparatus disclosed in the above prior art has the following problems.
First, it is necessary to set an optimal detection condition for detecting a workpiece that can be gripped. If the detection condition is not set properly, for example, contour extraction fails, and the overlapping state is erroneously recognized to reduce the success rate of the gripping operation. If such a situation occurs frequently, work efficiency will be poor.

Second, we need to construct a detection algorithm that takes into consideration the posture of all the work and disturbances. The work in the overlapping state has various positions, so the illumination is not uniform and the appearance is not constant. Also, as the disturbance, it is also necessary to consider the influence of extraneous matter such as oil on the surface of the work and disturbance light. It is not easy to devise a detection algorithm that covers all conditions, it requires many trial and error, and development is long-term.

A machine learning detection method has been proposed to solve these problems. In particular, examples of achieving recognition rates that surpass humanity in image recognition by deep learning have also been reported (ImageNet Large Scale Visual Recognition Challenge 2015). The reason for this is that it was possible to learn and learn a great many image data using the Internet and the like.

However, in the case of industrial products handled by robots, assembling devices, etc., it is difficult to collect many images in the manufacturing process, although a large amount of image data is required for machine learning. Therefore, it is an issue to collect many image data required for learning.

The present invention has been made to solve the above-mentioned problems, and its object is to require human trial and error and expert knowledge for selecting image features necessary for shape recognition of workpieces and selecting optimum conditions. Not to provide a work information processing apparatus by machine learning.

The present invention, in summary, is a workpiece information processing apparatus used for workpiece shape recognition, and a training data creation unit for creating training data necessary for learning based on workpiece design information, and a workpiece for learning training data A learning unit, and a data storage unit for storing training data and a learning result in the work learning unit.

Preferably, the work information processing apparatus further includes a work data creation unit that creates work data including a basic image of the work and coordinate data from the design information. The learning data creation unit changes the basic image to create training data.

More preferably, the learning data creation unit creates training data by performing at least one of translation, rotation, reduction, and / or enlargement on the basic image.

More preferably, the learning data creation unit changes at least one of the change of the color of the surface of the work, the change of the irradiation direction of the light irradiated to the work, and the change of the light intensity irradiated to the work with respect to the basic image. Create training data by performing one process.

Preferably, the workpiece information processing apparatus further includes a workpiece observation unit for photographing a workpiece. The training data includes the first image data created by the learning data creation unit and the second image data captured by the work observation unit.

Preferably, the design information of the workpiece includes two-dimensional CAD or three-dimensional CAD design information.

Preferably, the workpiece information processing apparatus further includes a workpiece recognition unit that receives an image of the workpiece and recognizes the shape of the workpiece. The work learning unit changes internal parameters of the work recognition unit by learning training data.

In another aspect, the present invention relates to a method of recognizing a work, which comprises: creating training data necessary for learning based on design information of the work; training training data; and training data and training data And storing the learning result of the step and receiving an image of the work, and using the learning result to recognize the shape of the work.

According to the present invention, it is possible to realize a work information processing apparatus which does not require human trial and error or expert knowledge for selecting an image feature necessary for shape recognition of a work and selecting an optimum condition.

It is a figure which shows the structural example of the workpiece | work information processing apparatus 1 of this Embodiment. It is a figure for demonstrating the training data which the learning data preparation part 3 produces. It is a figure which shows an example of training data. It is a figure which shows an example of the correspondence of a data name and the shape of a workpiece | work. It is a flow chart for explaining the 1st creation procedure of a picture for training. It is a figure which shows the structure used for the 2nd preparation procedure of the image for training. It is a flowchart for demonstrating the 2nd preparation procedure of the image for training. 5 is a diagram for explaining learning in a work learning unit 5. FIG. It is a flowchart for demonstrating a work learning procedure. It is a flowchart for demonstrating the workpiece | work identification process performed by a workpiece | work identification part.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following drawings, the same or corresponding parts are denoted by the same reference numerals, and the description thereof will not be repeated.

FIG. 1 is a diagram showing a configuration example of a work information processing apparatus 1 according to the present embodiment. Referring to FIG. 1, the work information processing apparatus 1 includes a work data creation unit 2, a learning data creation unit 3, a work observation unit 4, a work learning unit 5, a work recognition unit 6, and a data storage unit 7. And

The work data creation unit 2 creates work data as a reference of training data necessary for learning. The workpiece data creation unit 2 creates workpiece data based on the design information of the workpiece 8 including the material of the workpiece 8 and the processing method (surface property), and stores the workpiece data in the data storage unit 7. The design information of the work 8 includes two-dimensional CAD or three-dimensional CAD design information.

Although the substance of "work data" is an image, two-dimensional or three-dimensional dimensional information is stored together with the work data, and is stored in a commonly used two-dimensional or three-dimensional CAD data format. In addition, information such as material, processing method, surface quality and the like is also stored as necessary. As this format, DXF (Drawing Exchange Format), IGES (Initial Graphics Exchange Specification), STEP (Standard for the Exchange of Product model)
data) etc., and these can be used in the present embodiment.

The learning data creation unit 3 creates training data necessary for learning based on the work data created by the work data creation unit 2. As such training data, an image actually taken by a camera or the like is often used, but it is laborious to prepare a large number of such images with a machine tool or the like. Therefore, in the present embodiment, the learning data creation unit 3 automatically generates a large number of training data from the original work data. FIG. 2 is a figure for demonstrating the training data which the learning data creation part 3 produces. The learning data creation unit 3 modifies the basic image 100 to create training data. As shown in FIG. 2, the training data creation unit 3 performs

training data

100 A, 100 B by performing at least one of translation, rotation, reduction and enlargement on the basic image 100 according to a program. Create

Preferably, as shown in FIG. 2, the learning data creation unit 3 causes the basic image 100 to change the color of the surface of the work, change the irradiation direction of light to be irradiated to the work, and irradiate the work with the program.

Training data

100C and 100D are created by performing at least one process of changing the light intensity.

As described above, the learning data creation unit 3 artificially creates various training data from CAD data, not from actual photographed images. The learning data creation unit 3 is a program for performing at least one of the above-mentioned translation, rotation, reduction, and enlargement, the change of the color of the work surface, the change of the direction of the light to be irradiated, and the intensity. It is a combination of programs that perform at least one process of change, and the order of processes can be arbitrarily changed.

FIG. 3 is a diagram showing an example of training data. As shown in FIG. 3, the training data is an image in which sample points (pixels) are gathered. The training data includes the contour W1 of the work.

The training data consists of three-dimensional coordinates (X, Y, Z) of each sample point on the work surface, the color (R, G, B) of the work, and the brightness I, which are stored in an array and associated with data names It is stored in the storage unit 7. The data name indicates the type of created data, and may be numbered sequentially from 1 or may be represented by characters such as A and ABC.

FIG. 4 is a diagram showing an example of correspondence between data names and shapes of workpieces. The data names "1" to "5" correspond to five types of workpiece shapes. Specifically, a square work corresponds to the data name "1", a circular work corresponds to the data name "2", and a parallelogram work corresponds to the data name "3" The data name "4" corresponds to a triangular work, and the data name "5" corresponds to a double circle work. Although FIG. 4 shows a simple two-dimensional shape as a shape example for the sake of simplicity, in actuality, the shape of the workpiece may be three-dimensional.

In addition, the condition when training data is created in relation to the data name, for example, the type of work, parallel movement amount, rotation angle, scale value, color, light irradiation direction, light intensity, etc. It is stored in section 7.

Referring back to FIG. 1 again, the workpiece observation unit 4 outputs three-dimensional coordinates (X, Y, Z), colors (R, G, B) and brightness I of each sample point on the workpiece surface. In the configuration example of FIG. 1, the workpiece observation unit 4 includes two cameras for photographing the workpiece 8 and obtains three-dimensional coordinates (X, Y, Z) of each sample point on the workpiece surface by stereo measurement. These are color cameras and can output the color (R, G, B) and brightness I of each sample point on the surface of the workpiece.

The work observation unit 4 is also connected to the work learning unit 5 and the work recognition unit 6. The work learning unit 5 and the work recognition unit 6 execute the learning process and the recognition process using data (X, Y, Z), (R, G, B), and I, respectively.

The work learning unit 5 performs learning using the training data created by the learning data creation unit 3. As a learning method, supervised learning methods such as support vector machine and back-provocation (error back propagation method), unsupervised learning methods such as auto encoder (self-coder), k-means method, principal component analysis, and Use machine learning method that combines them.

The workpiece recognition unit 6 performs workpiece recognition processing using the output data of the workpiece observation unit 4 and the learning result of the workpiece learning unit 5. The workpiece recognition unit 6 uses a recognition model such as a neural network or support vector machine corresponding to the machine learning method, and based on the learning result of the workpiece learning unit 5, each specimen of the workpiece surface output from the workpiece observation unit 4 The recognition processing is performed using the three-dimensional coordinates (X, Y, Z) of the point, the colors (R, G, B), and the brightness I as inputs.

Each of the work data creation unit 2, the learning data creation unit 3, the work learning unit 5, and the work recognition unit 6 includes a central processing unit (CPU), a random access memory (RAM), a read only memory (ROM), etc. Control of each component is performed according to processing. The functions of two or more of the work data creation unit 2, the learning data creation unit 3, the work learning unit 5, and the work recognition unit 6 may be processed by one CPU. The data storage unit 7 is, for example, an auxiliary storage device such as a hard disk drive or a solid state drive. The data storage unit 7 is a program executed by each of the work data generation unit 2, the learning data generation unit 3, the work learning unit 5, and the work recognition unit 6, and the work data generation unit 2 and the learning data generation unit 3. The data and the learning result of the work learning unit 5 are stored.

[Preparation of images for training]
The work learning unit 5 changes the internal parameter of the work recognition unit 6 by combining the first image data prepared in the first preparation procedure described below and the second image data prepared in the second preparation procedure. Execute learning process. The internal parameters correspond to, for example, weights and biases between layers from the input layer to the output layer in the neural network. That is, the training data learned by the work learning unit 5 includes the first image data and the second image data. The first preparation procedure is a procedure for generating a large number of first image data for training from design data (CAD data). The second preparation procedure is a procedure in which an image of an actual workpiece acquired by the workpiece observation unit 4 is used as second image data for training.

FIG. 5 is a flowchart for explaining a first procedure for creating a training image. The first creation procedure is executed in the learning data creation unit 3 of FIG. Referring to FIGS. 1 and 5, in step S1, the learning data creation unit 3 reads CAD data (original design data) created by the work data creation unit 2. In step S2, the learning data creation unit 3 inquires to which data name the read CAD data corresponds, and the user selects a data name corresponding to the CAD data.

Subsequently, in step S3, the learning data creation unit 3 creates a correct answer corresponding to the data name input by the user. The correct answer corresponds to the recognition result output from the workpiece recognition unit 6. For example, when the probability corresponding to data names "1" to "5" in FIG. 4 is output as recognition results P1 to P5 from work recognition unit 6, the correct answer corresponding to data name "3" is (P1 = 0 , P2 = 0, P3 = 1, P4 = 0, P5 = 0).

Subsequently, the learning data creation unit 3 determines the change range of the parameter in step S4. The parameters include the amount of parallel movement, the angle of rotation, the reduction or enlargement ratio, the color of the work, the irradiation direction and intensity of light, etc. The learning data creation unit 3 determines the change range of the parameter according to the input from the user. In addition, the learning data creation unit 3 may set the change range of the parameter to a value determined in advance as a standard.

Subsequently, in step S5, the learning data creation unit 3 creates an image while changing the parameter within the change range determined in step S4. For example, the amount of parallel movement, rotation angle, magnification, color work, the irradiation direction of the light, the six parameters of the irradiation intensity of light, when changing the respective five levels within the change area, 5 ⁶ image is generated Be done.

In step S6, the learning data creation unit 3 stores the generated image in the data storage unit 7 in association with the parameter and the data name.

FIG. 6 is a diagram showing a configuration used for a second procedure for creating a training image.
In the second preparation procedure, an image taken by the work observation unit 4 is used. Only one of the training image prepared in the first preparation procedure or the training image prepared in the second preparation procedure may be used, but the training image prepared in the first preparation procedure and the second preparation procedure The recognition rate can be further improved by combining the prepared training image. The training data created by the learning data creation unit 3 is reinforced by causing the work observation unit 4 to simultaneously learn training data obtained by observing an actual target work. The image generated in the first preparation procedure is nothing arranged around the work image. On the other hand, in the image created in the second preparation procedure, another workpiece may be shown around the workpiece. In the later learning process, the work learning unit 5 executes learning so as to recognize the work with the surrounding image. Therefore, even when a plurality of works 8 are contained in a container and a part thereof is overlapped, one work can be identified and recognized.

The workpiece observation unit 4 outputs three-dimensional coordinates (X, Y, Z), colors (R, G, B) and brightness I of each sample point on the workpiece surface. Since these are not associated with data names, they will be associated with data names at the time of acquisition.

FIG. 7 is a flowchart for describing a second procedure for creating a training image. The work learning unit 5 acquires the image captured from the work observing unit 4 in step S21. The work learning unit 5 displays the acquired acquired data 12 on the monitor 9 connected to the work learning unit 5. Further, in step S22, the user selects one of the data names 13 using the keyboard 10 and the mouse 11 connected to the work learning unit 5, thereby acquiring the data name corresponding to the acquired data 12. At this time, the acquired data 12 displayed on the monitor 9 and the selected data name 13 are associated with each other and stored in the data storage unit 7. Subsequently, in step S23, the learning data creation unit 3 creates a correct answer corresponding to the data name 13 input by the user. The correct answer corresponds to the recognition result output from the workpiece recognition unit 6. For example, when the probability corresponding to data names "1" to "5" in FIG. 4 is output as recognition results P1 to P5 from work recognition unit 6, the correct answer corresponding to data name "3" is (P1 = 0 , P2 = 0, P3 = 1, P4 = 0, P5 = 0).

In FIG. 6, the monitor 9 is connected to the work learning unit 5, and the work learning unit 5 executes the above-described second preparation procedure. However, the monitor 9 is connected to the work recognition unit 6 and connected to the work recognition unit 6. The second recognition procedure described above may be performed by the work recognition unit 6 using the keyboard 10 and the mouse 11. In this case, the workpiece recognition unit 6 and the data storage unit 7 are connected so that data can be stored.

In steps S24 to S27, in the above association, the user inputs the position, the rotation angle, the scale, and the like of the work using the keyboard 10 and the mouse 11 while looking at the acquired data 12 displayed on the monitor 9.

In step S24, the work learning unit 5 displays the acquired image on the monitor 9. Subsequently, in step S25, the work learning unit 5 receives specification of the center position of the work in the image from the user, and calculates the amount of parallel movement. The amount of parallel movement is given as coordinate values in the horizontal and vertical directions with the upper left end of the image as the origin.

Furthermore, in step S26, the work learning unit 5 calculates the rotation angle based on two points of the work in the image designated by the user. For example, if the work is a square, the user specifies two diagonal corners. Thus, the work learning unit 5 can recognize the rotation angle. In step S27, when the user designates an outer shape on the image, the work learning unit 5 can calculate the magnification of the work captured in the image.

Then, in step S28, the input position, rotation angle, scale, and the like are stored in the data storage unit 7 in association with the acquired data 12.

After the training data is prepared in the first creation procedure or the second creation procedure described above, the work learning unit 5 performs learning in the procedure described below.

[Learning using training images]
FIG. 8 is a diagram for explaining learning in the work learning unit 5. FIG. 9 is a flowchart for explaining the work learning procedure executed by the work learning unit. The following description is given taking as an example the case of learning the data name "3".

The learning data creation unit 3 stores data names in association with the training data. In step S31, the work learning unit 5 together with the training data from the data storage unit 7 as teacher data of the data name “3” (P1 = 0, P2 = 0, P3 = 1, P4 = 0, P5 = 0) Get

In step S32, the work learning unit 5 initializes a variable n for counting the number of images to one. In step S33, it is determined whether or not the variable n is equal to or less than the number of images. In step S33, when the variable n is equal to or less than the number of images, the processes of steps S34 to S37 are performed. In step S33, when the variable n exceeds the image, the processing in steps S34 to S37 is not executed, and the process proceeds to step S38.

In step S34, the work learning unit 5 inputs training data to the work recognition unit 6. Then, in step S35, the work learning unit 5 acquires the recognition result of the work recognition unit 6.

In step S36, the work learning unit 5 corrects the internal parameters of the work recognition unit 6 so that the output result of the work recognition unit 6 matches the teacher data. The recognition is performed again using the changed internal parameter, and if the output of the workpiece recognition unit 6 does not match the teacher data, the workpiece learning unit 5 corrects the internal parameter again. In step S36, the work learning unit 5 repeatedly performs internal parameter correction and recognition processing, and brings the output of the work recognition unit 6 closer to teacher data.

In FIG. 8, the recognition results P1 to P5 indicate the probabilities corresponding to the data names "1" to "5", respectively. (P1 = 1, P2 = 0, P3 = 0.5, P4 = 1, P5 = 0) indicates the recognition result of the training data in the work recognition unit 6 before adjusting the internal parameter. In the teacher data (P1 = 0, P2 = 0, P3 = 1, P4 = 0, P5 = 0) shown in FIG. 8, the probability of the data name “3” is 1 and the data names “1” and “2” "," "4", "5" indicates the recognition result that the probability is zero. In step S36, the work learning unit 5 performs step S36 so that the result of recognition of the training data approaches teacher data (P1 = 0, P2 = 0, P3 = 1, P4 = 0, P5 = 0). Execute learning processing to adjust internal parameters.

Following step S36, the variable n is incremented in step S37, and the determination process of step S33 is performed again. If the variable n exceeds the number of training images in the determination process of step S33, the process proceeds to step S38, and the learning result is stored. Specifically, in step S38, the work learning unit 5 fixes the internal parameter of the work recognition unit 6 that has been adjusted, and ends the learning.

In the above description, the procedure for learning the data name has been exemplified. However, in addition to the data name, the work learning unit 5 translates the parallel movement amount, rotation angle, and scale of the work corresponding to the acquired data when the learning data is created. Etc. may be given as teacher data, and these may be directly learned. In this case, as a result of the recognition process of the workpiece recognition unit 6, the parallel movement amount, the rotation angle, the scale, etc. are output from the workpiece recognition unit 6. The amount of parallel movement is given as coordinate values in the horizontal and vertical directions with the upper left end of the image as the origin.

[Work identification process]
The workpiece recognition unit 6 whose internal parameters have been adjusted in the learning process can identify which data name the workpiece photographed by the workpiece observation unit 4 corresponds to.

FIG. 10 is a flowchart for explaining the workpiece identification process performed by the workpiece identification unit. First, in step S41, the workpiece recognition unit 6 acquires an image captured by the workpiece observation unit 4. Subsequently, in step S42, the work recognition unit 6 executes the recognition process using the internal parameters adjusted by the work learning unit 5. Then, in step S43, the workpiece recognition unit 6 outputs the recognition result.

The work recognition method described in the above flow chart will be summarized and described. The work recognition method according to the present embodiment includes steps (S1 to S6) of creating training data necessary for learning based on design information of the work, and steps (S31 to S38) of learning training data. Receiving an image of the workpiece and recognizing the shape of the workpiece using the learning result (S41 to S43).

As described above, since a large number of training data are created based on design information of a workpiece, it is possible to reduce the trouble of actually photographing the workpiece with a camera and preparing a large number of training images for learning.

Further, in machine learning, since the correct answer is known from the data itself including the disturbance, even in the case of erroneous recognition, there is no need to change the processing content in the above-described work recognition method. When new erroneously recognized image data is generated, only learning of the data itself is required, so that there is also an advantage that no specialized knowledge of image processing is required.

Such a recognition result can be used, for example, to recognize which of the five shapes shown in FIG. 4 contains the workpiece in the container for containing the workpiece. The posture of the workpiece can also be detected by the amount of parallel movement, the angle of rotation, the scale, etc. included in the recognition result.

The image feature necessary for detecting the posture of the work is learned by the neural network itself consisting of the work learning unit 5 and the work recognition unit 6, so that the number of steps for examining the detection algorithm can be reduced. In addition, since it is not necessary to search for or adjust the optimum detection conditions, complicated operations can be omitted, and the number of steps for constructing a recognition device can be reduced.

It should be understood that the embodiments disclosed herein are illustrative and non-restrictive in every respect. The scope of the present invention is shown not by the above description of the embodiment but by the scope of claims, and is intended to include all modifications within the scope and meaning equivalent to the scope of claims.

1 Work information processing apparatus, 2 Work data creation unit, 3 Learning data creation unit, 4 Work observation unit, 5 Work learning unit, 6 Work recognition unit, 7 Data storage unit, 9 monitors, 10 keyboards, 11 mice.

Claims

A workpiece information processing apparatus used for workpiece shape recognition,
A learning data creation unit that creates training data necessary for learning based on design information of the work;
A work learning unit for learning the training data;
A work information processing apparatus comprising: a data storage unit for storing the training data and the learning result in the work learning unit.
It further comprises a work data creation unit that creates work data including the basic image of the work and coordinate data from the design information,
The work information processing apparatus according to claim 1, wherein the learning data creation unit creates the training data by changing the basic image.
The work according to claim 2, wherein the training data creation unit creates the training data by performing at least one of translation, rotation, reduction, and / or enlargement on the basic image. Information processing device.
The learning data creation unit is at least one of a change of a color of a surface of the work, a change of an irradiation direction of light irradiated to the work, and a change of intensity of light irradiated to the work with respect to the basic image. The work information processing apparatus according to claim 2, wherein the training data is created by performing any one process.
The apparatus further comprises a workpiece observation unit for photographing the workpiece,
The training data is
First image data created by the learning data creation unit;
The work information processing apparatus according to any one of claims 1 to 4, including the second image data captured by the work observing unit.
The workpiece information processing apparatus according to any one of claims 1 to 5, wherein the design information of the workpiece includes two-dimensional CAD or three-dimensional CAD design information.
The apparatus further comprises a workpiece recognition unit that receives an image of the workpiece and recognizes the shape of the workpiece.
The work information processing apparatus according to any one of claims 1 to 6, wherein the work learning unit changes an internal parameter of the work recognition unit by learning the training data.
Creating training data necessary for learning based on the design information of the work;
Learning the training data;
Receiving an image of the workpiece and recognizing a shape of the workpiece using a learning result.