Monitoring device

Abstract

To monitor a molding state in real time by imaging a mold parting surface every time.SOLUTION: A monitoring device for estimating a molding state in a molding process of a resin molded body includes: imaging means for imaging a mold parting surface capable of forming a mold cavity between a fixed mold and a movable mold opposed to the fixed mold to obtain a captured image; extraction means for extracting a feature amount of the mold parting surface from the captured image; state estimation means for estimating the molding state based on the feature amount; and notification means for notifying it when the molding state reaches a predetermined state.SELECTED DRAWING: Figure 10

Description

Technical Field

[0001] The present invention relates to a monitoring device.

Background Art

[0002] An abnormality detection device for detecting an abnormality in an injection molding machine that manufactures molded products in each of a plurality of cavities formed inside a mold composed of a fixed mold and a movable mold movable with respect to the fixed mold, and filling each cavity with resin, the device includes: an image acquisition unit that acquires an image of at least one mold surface of the fixed mold and the movable mold during molding as an inspection target image; an abnormality detection unit that detects the cavity in which an abnormality has occurred based on a comparison between the inspection target image and a reference image; and a signal output unit that outputs, in association with each other, identification information of the cavity in which an abnormality has been detected by the abnormality detection unit and an abnormality signal. (Patent Document 1)

[0003] An input data acquisition unit that acquires input data including at least any molding conditions including the type of resin, the type of additive, the blending ratio of the additive, and the temperature of the resin in the molding of an arbitrary molded product by an arbitrary injection molding machine, and state information indicating the wear amount of the mold before molding under the molding conditions; a label acquisition unit that acquires label data indicating the state information of the mold after molding under the molding conditions included in the input data; and a learning unit that performs supervised learning using the input data acquired by the input data acquisition unit and the label data acquired by the label acquisition unit to generate a learned model. (Patent Document 2)

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] The present invention captures the mold split surface each time and monitors the molding state in real time.

Means for Solving the Problems

[0006] In order to solve the above problems, the monitoring device according to claim 1 is a monitoring device for estimating a molding state in the molding process of a resin molded body, imaging means for capturing an imaging image by imaging a mold split surface capable of forming a mold cavity between a fixed mold and a movable mold opposed to the fixed mold; extracting means for extracting a feature amount of the imaging image; image processing means for estimating the molding state based on the feature amount; notification means for notifying when the molding state reaches a predetermined state, characterized by that.

[0007] The invention according to claim 2 is the monitoring device according to claim 1, wherein the imaging image is an image captured each time before closing the fixed mold and the movable mold in one molding cycle. characterized by that.

[0008] The invention according to claim 3 is the monitoring device according to claim 2, wherein the imaging image is a frame image extracted at regular time intervals from a moving image of the mold split surface. characterized by that.

[0009] The invention according to claim 4 is the monitoring device according to claim 1, wherein the feature amount is a black image obtained by performing binarization processing using a threshold value preset in the imaging image to convert the entire image into a black and white image. characterized by that.

[0010] The invention according to claim 5 is the monitoring device according to claim 1, wherein the feature amount is a contour line extracted from the captured image. This is the gist of the invention.

[0011] The invention according to claim 6 is the monitoring device according to claim 1, wherein the feature amount is a divided region obtained by dividing the captured image into regions with a preset density and / or color. This is the gist of the invention.

[0012] The invention according to claim 7 is the monitoring device according to any one of claims 1 to 6, wherein the state estimation means estimates the formed state based on the difference between the feature amount and the feature amount extracted from a reference image which is the captured image captured first. This is the gist of the invention.

[0013] The invention according to claim 8 is the monitoring device according to any one of claims 1 to 6, wherein the state estimation means estimates the formed state from the feature amount extracted by the extraction means using a learned model obtained by machine learning for estimating the formed state from the captured image. This is the gist of the invention.

[0014] The invention according to claim 9 is the monitoring device according to any one of claims 1 to 6, wherein the state estimation means calculates the similarity between the feature amount and a template image prepared in advance, and estimates the formed state based on the similarity. This is the gist of the invention.

[0015] The invention according to claim 10 is the monitoring device according to claim 1, wherein the imaging means is attached to the resin molded body take-out device, and images the mold parting surface when the fixed mold and the movable mold are open. This is the gist of the invention.

Advantages of the Invention

[0016] According to the invention described in claim 1, it is possible to image the mold split surface and monitor the molding state in real time.

[0017] According to the invention described in claim 2, it is possible to acquire an image of the mold split surface in real time.

[0018] According to the invention described in claim 3, it is possible to facilitate image processing for estimating the molding state.

[0019] According to the invention described in claim 4, it is possible to reliably detect resin residues including dirt, burns, burrs, etc. on the mold split surface and partial resin residues in the cavity.

[0020] According to the invention described in claim 5, it is possible to reliably detect dirt, burns, burrs, etc. on the mold split surface.

[0021] According to the invention described in claim 6, it is possible to reliably detect the area, shape, and number of dirt, burns, etc. on the mold split surface.

[0022] According to the invention described in claim 7, it is possible to detect resin residues including dirt, burns, burrs, etc. on the mold split surface and partial resin residues in the cavity, and estimate the molding state.

[0023] According to the invention described in claim 8, it is possible to detect resin residues including dirt, burns, burrs, etc. on the mold split surface and partial resin residues in the cavity, and estimate the molding state.

[0024] According to the invention described in claim 9, it is possible to detect resin residues including dirt, burns, burrs, etc. on the mold split surface and partial resin residues in the cavity, and estimate the molding state.

[0025] According to the invention described in claim 10, it is possible to image the mold split surface in real time for each molding cycle.

Brief Description of the Drawings

[0026]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Mode for Carrying Out the Invention

[0027] Next, specific examples of embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to the following embodiments. Note that in the description using the following drawings, the drawings are schematic, and it should be noted that the ratios of each dimension, etc. are different from the actual ones, and illustrations other than the members necessary for the description are appropriately omitted for ease of understanding.

[0028] (1) Injection molding system First, an injection molding system 100 to which the monitoring device 1 according to the present embodiment is applied will be described with reference to the drawings. FIG. 2 is a schematic diagram showing a configuration example of an injection molding system 100 to which the monitoring device 1 according to the present embodiment is applied, FIG. 3 is a schematic diagram showing a configuration example of a mold K, FIG. 3(a) is a schematic cross-sectional view showing the configuration of the main part of the mold K, and (b) is a schematic cross-sectional view showing the main part of the mold K in an open state. The injection molding system 100 includes an injection molding machine 110, a take-out device 120 for taking out a resin molded body S from the mold K, and a monitoring device 1 for estimating the molding state in the molding process of the resin molded body S.

[0029] The injection molding machine 110 includes at least a mold clamping unit 111, an injection unit 112, a mold K composed of a fixed-side mold K1 and a movable-side mold K2, and a control unit 113. The clamping unit 111 performs clamping by moving the movable mold K2 closer to the fixed mold K1. The injection unit 112 fills the cavity CA (see Fig. 3(a)) formed between the clamped movable mold K2 and the fixed mold K1 with the molten resin used for molding the resin molded body S (not shown). After the molten resin filled in the cavity CA solidifies, the clamping unit 111 moves the movable mold K2 away from the fixed mold K1 to perform mold opening (a series of operations for obtaining the resin molded body S from the start of the clamping process through the mold opening process to the start of the next clamping process is also called a "shot" or a "molding cycle").

[0030] The mold K is composed of the fixed mold K1 and the movable mold K2 joined at the mold split surface PL, and a cavity CA for filling resin is formed between the fixed mold K1 and the movable mold K2. The fixed mold K1 consists of a fixed insert K11 in which a cavity surface constituting a part of the cavity CA and a part of the mold split surface PL are formed, and a gas vent insert K12 having a gas vent groove GV1 formed at one end. The movable mold K2 has a movable insert K21 having a desired core K22 on the cavity surface of the fixed insert K11. A gas vent groove GV2 that communicates with one end of the gas vent groove GV1 and exhausts the gas generated in the cavity CA to the outside of the mold K is formed on the mold split surface PL of the movable insert K21 (hereinafter, the gas vent groove GV1 and the gas vent groove GV2 may be referred to as the gas vent part GV without distinction). Note that the fixed mold K1 and the movable mold K2 shown in Fig. 3 are examples, and the configurations of the cavity CA and the gas vent part GV are not limited thereto. For example, the cavity CA and the gas vent part GV may be formed in the fixed mold K1, and the movable mold K2 may be configured to consist only of a flat mold split surface PL.

[0031] The take-out device 120 includes a slide guide 121, moving parts 122 and 123, and a take-out part 124. The slide guide 121 extends in a horizontal direction intersecting the moving direction of the movable mold K2, and guides the horizontal movement of the moving part 122 (see arrow R2 in Fig. 2). The moving part 123 is vertically movable with respect to the moving part 122. The moving part 122 horizontally moves the moving part 123 and the take-out part 124 under the guidance of the slide guide 121. The take-out part 124 is horizontally movable between the take-out position above the movable mold K2 and a position outside the injection molding machine 110 where the resin molded body S is released away from the movable mold K2 (open position). Note that the take-out device 120 shown in Fig. 2 is an example, and it is not limited to the specific configuration shown in Fig. 2 as long as it has the function of taking out the resin molded body S from the opened mold K.

[0032] The monitoring device 1 includes an imaging unit 10 that captures an imaging image by imaging a mold parting surface PL capable of forming a cavity CA between a fixed mold K1 and a movable mold K2, an extraction unit 20 that extracts a feature amount F of the captured imaging image, a state estimation unit 30 that estimates a molding state based on the feature amount F, a notification unit 40 that notifies that the molding state has reached a predetermined state when this occurs, and a storage unit 50 that stores image data and the feature amount F, and monitors the molding state in real time.

[0033] (2) Overall configuration of the monitoring device Fig. 1 is a functional block diagram showing an example of the monitoring device 1 according to the present embodiment. The monitoring device 1 includes an imaging unit 10 that captures an image of a mold split surface PL including a cavity CA between a fixed-side mold K1 and a movable-side mold K2 facing the fixed-side mold K1 to acquire image data, an extraction unit 20 that extracts a feature amount F of the captured imaging image, a state estimation unit 30 that estimates a molding state based on the feature amount F extracted by the extraction unit 20, and a notification unit 40 that notifies when the molding state reaches a predetermined state, and monitors the molding state in real time. Here, examples of the molding state include dirt and burning of the gas vent portion GV on the mold split surface PL including the cavity CA, resin residue on the mold split surface PL including burrs, etc., and partial resin residue of the resin molded body S in the cavity CA.

[0034] In the monitoring device 1, the operation control and processing control of the imaging unit 10, the extraction unit 20, the state estimation unit 30, and the notification unit 40 are realized by a general computer in which a processor including a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory) executes a predetermined program. In the present embodiment, it can be executed using a Raspberry Pi (registered trademark), a single-board computer equipped with an ARM processor, but it is not particularly limited to the Raspberry Pi.

[0035] (Imaging unit) The imaging unit 10 captures an image of the mold split surface PL in a state where the fixed-side mold K1 and the movable-side mold K2 are open. The imaging unit 10 can be fixedly installed on, for example, the take-out device 120. In the present embodiment, as shown in FIG. 2, it is installed on the take-out portion 124 of the take-out device 120 so as to face the open mold split surface PL.

[0036] The imaging unit 10 is composed of a camera equipped with a CCD image sensor as an imaging device and having an imaging condition adjustment function for adjusting optical conditions such as the angle of view and exposure, and captures still images or moving images. In the case of moving image capture, frame images extracted at regular time intervals are cut out. As a result, even when capturing as a moving image, only the frame images whose image analysis target is a still image are obtained, facilitating image processing for estimating the molding state. As the camera, as long as it has the resolution and other requirements necessary for the extraction unit 20 to extract the feature quantity F and has the function of sending the captured image as digital data to the extraction unit 20, it may be a dedicated imaging device, or for example, a general-purpose imaging device that can be generally obtained as a peripheral device of a personal computer such as a USB camera, a WEB camera, or a network camera.

[0037] In addition, in this embodiment, there is one imaging unit 10, but a plurality of imaging units 10 may be installed. The mold parting surfaces PL of the fixed-side mold K1 and the movable-side mold K2 opened by the plurality of imaging units 10 may be imaged separately (or partially overlapped), or the entire mold parting surface PL may be imaged by each of the plurality of imaging units 10 from different angles. The captured images captured by the plurality of imaging units 10 may be subjected to synthesis processing so that, for example, the entire mold parting surface PL becomes one image.

[0038] (Extraction unit) The extraction unit 20 acquires the captured image data (hereinafter simply referred to as image data) captured by the imaging unit 10. The extraction unit 20 acquires the image data from the imaging unit 10, for example, by wireless communication with the imaging unit 10. Examples of wireless communication methods include Wi-Fi (registered trademark), Bluetooth (registered trademark), and ZigBee (registered trademark). The extraction unit 20 may also acquire the image data from the imaging unit 10 by wired communication with the imaging unit 10.

[0039] The extraction unit 20 may obtain image data from the imaging unit 10 by communicating with the imaging unit 10 via a network, for example. The network includes, for example, the Internet and a LAN (Local Area Network). Further, the network may include a mobile communication network. The mobile communication network may conform to, for example, a communication method such as 3G (3rd Generation), LTE (Long Term Evolution), 5G (5th Generation), or a communication method subsequent to 6G (6th Generation).

[0040] The extraction unit 20 extracts a feature amount F of the image data acquired from the imaging unit 10. The feature amount F extracted by the extraction unit 20 is stored in the storage unit 50 together with the image data. Specifically, the extraction unit 20 performs binarization processing using a preset threshold Th1 on the acquired image data to convert the entire image into a black-and-white image and extract a black image. Here, the black image as the feature amount F indicates the occurrence of resin residues such as stains, burns, and burrs on the mold parting surface PL, and it becomes possible to estimate the molding state based on such a black image.

[0041] In the binarization processing, for each pixel in the image data, the pixel value is set to "0" or "1" with a predetermined threshold Th1 as the boundary. For example, when the luminance value is equal to or greater than the predetermined threshold Th1, the pixel value of the pixel is set to "1" (for example, white), and when it is less than the threshold Th1, the pixel value of the pixel is set to "0" (for example, black).

[0042] FIG. 4(a) shows image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 after use. FIGS. 4(b), (c), and (d) are diagrams showing examples of image data obtained by performing binarization processing while changing the threshold Th1 to Th11, Th12, and Th13, respectively. As shown in FIGS. 4(b), (c), and (d), by changing the threshold Th1 for binarization processing, the range (region) where the pixel value becomes "0" (for example, black) changes. That is, by appropriately setting the threshold Th1 for binarization processing, the degree of contamination of the gas vent portion GV on the mold split surface PL as the shaped state to be estimated can be set.

[0043] As the threshold Th1 for binarization processing, the luminance value of the mold split surface PL of the unused mold K is used. The luminance value of the mold split surface PL of the unused mold K may be a previously measured value or a value estimated from a reference image that is an image of the mold split surface PL of the mold K imaged at the beginning of molding.

[0044] FIG. 5(a) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 immediately after the start of molding and image data obtained by performing binarization processing. FIG. 5(b) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 during molding and image data obtained by performing binarization processing. FIG. 5(c) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 immediately before the end of molding and image data obtained by performing binarization processing. In the present embodiment, taking the contamination (deposit in which volatile components and gas components generated from the molten resin are liquefied) of the gas vent portion GV on the mold split surface PL as one of the detection targets, a pixel value greater than the threshold Th1 for binarization processing is set to "1", and a pixel value equal to or less than the threshold Th1 for binarization processing is set to "0". As a result, the pixel value of the gas vent portion GV on the mold split surface PL of the mold K immediately after the start of molding becomes "1" (for example, white), and a part of the gas vent portion GV on the mold split surface PL of the mold K during molding (for example, after 150 shots) and immediately before the end of molding (for example, after 300 shots) becomes a pixel value of "0" (for example, black).

[0045] As the image data to be binarized, the same applies to an image obtained by imaging burning on the mold parting surface PL, an image obtained by imaging resin residue on the mold parting surface PL including burrs, etc., and an image obtained by imaging a part of the resin residue of the resin molded body S in the cavity CA. Binarization processing is performed using a predetermined threshold value Th1 to detect burning, burrs, and partial resin residue, and the entire image is converted into a black-and-white image.

[0046] (Modification Example 1) The extraction unit 20 may extract a contour line CT from the captured image as the feature amount F. Specifically, the extraction unit 20 extracts at least one contour line CT by performing edge detection on the image data acquired from the imaging unit 10. Edge detection of the image data can be detected, for example, as a change in luminance value in the image data, but edge detection is a general process performed in image processing, and its details are omitted.

[0047] FIG. 6(a) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold parting surface PL of the movable mold K2 during molding and image data obtained by extracting the contour line CT, and FIG. 6(b) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold parting surface PL of the movable mold K2 immediately before the end of molding and image data obtained by extracting the contour line CT.

[0048] In the present embodiment, the feature amount F extracted as the contour line CT in the captured image indicates dirt (deposit in which volatile components and gas components generated from the molten resin are liquefied, etc.) of the gas vent portion GV on the mold parting surface PL, and it becomes possible to estimate the molding state based on such a contour line CT.

[0049] (Modification Example 2) The extraction unit 20 may extract, as the feature quantity F, a divided region RG that is divided into regions with a preset density and / or color from the captured image. Specifically, the extraction unit 20 divides and cuts out a region of a preset specific color from the image data acquired from the imaging unit 10, and divides and cuts out a region of a preset shading value, thereby extracting at least one divided region RG. Here, examples of the preset specific color include colors indicating dirt on the gas vent portion GV on the mold split surface PL (deposits formed by liquefaction of volatile components and gas components generated from the molten resin), and examples of the preset shading value include a density difference indicating the thickness of the deposit on the gas vent portion GV on the mold split surface PL.

[0050] FIG. 7(a) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 during molding and image data obtained by extracting the divided region RG, and FIG. 7(b) is a diagram showing an example of image data obtained by imaging the gas vent portion GV on the mold split surface PL of the movable mold K2 immediately before the end of molding and image data obtained by extracting the divided region RG.

[0051] In the present embodiment, the feature quantity F extracted as the divided region RG in the captured image indicates the degree of dirt (deposits formed by liquefaction of volatile components and gas components generated from the molten resin) on the gas vent portion GV on the mold split surface PL, the thickness of the deposit, etc., and it becomes possible to estimate the molding state based on such a divided region RG.

[0052] (State estimation unit) The state estimation unit 30 estimates the molding state based on the difference between the feature quantity F extracted by the extraction unit 20 and the feature quantity F extracted from the reference image, which is the first captured image. When the feature quantity F extracted by the extraction unit 20 is a black image when the entire image is converted into a black-and-white image by performing binarization processing using a preset threshold Th1 on the acquired image data, the ratio (black ratio BR) of the pixel value in the binarized image data being "0" (black) to the binarized reference image is calculated, and the molding state is estimated. When the feature amount F extracted by the extraction unit 20 is the contour line CT in the captured image, the forming state is estimated based on the change in the contour line CT with respect to the reference image. When the feature amount F extracted by the extraction unit 20 is the divided region RG divided into regions with a preset density and / or color in the captured image, the forming state is estimated based on the change in the divided region RG with respect to the reference image.

[0053] (Modification Example 1) The state estimation unit 30 may estimate the forming state from the feature amount F extracted by the extraction unit 20 using a learned model obtained by machine learning for estimating the forming state from the image data. In the estimation of the parameter representing the forming state, a plurality of feature amounts F extracted by the extraction unit 20 are used as inputs to the learning model. The state estimation unit 30 performs machine learning of a learning model that outputs an estimated value of the parameter representing the forming state from the feature amount F of the input image data.

[0054] The learning model (machine learning model) according to the present embodiment is a model that is trained using learning teacher data (a set of known input data and correct answer data) and enables prediction of future outputs. The teacher data can be obtained in test molding performed as a preliminary investigation of injection molding. For example, in test molding, injection molding is performed while changing the molding conditions, and the image data of the mold parting surface PL is acquired by the imaging unit 10. After the test molding, the stain on the gas vent part GV on the mold parting surface PL obtained is detected by, for example, surface observation. The feature amount F is extracted from the image data of the gas vent part GV thus obtained by the extraction unit 20 and can be used as the input data of the teacher data. As the machine learning method, various algorithms can be used depending on the purpose and conditions. For example, a neural network model based on deep learning can be used, but it is not limited to this.

[0055] (Modification Example 2) The state estimation unit 30 may calculate the similarity between the feature amount F extracted by the extraction unit 20 and a template image prepared in advance, and estimate the forming state based on the similarity. The memory unit 50 stores in advance a template image which is a captured image of the gas vent portion GV of the mold parting surface PL in a state where deposits or the like in which volatile components and gas components generated from the molten resin by molding have liquefied are adhered and soiled. For example, as shown in FIG. 8, image data A (at the start of molding), image data B (after 200 shots), and image data C (after 300 shots) obtained by imaging in advance the gas vent portion GV on the mold parting surface PL of the movable mold K2 for each molding shot are stored in the memory unit 50 as template images. The state estimation unit 30 reads out the template image stored in advance from the memory unit 50, and calculates the degree of similarity between the read template image and the feature amount F extracted by the extraction unit 20 from the image data captured by the imaging unit 10. If the calculated degree of similarity is equal to or greater than a predetermined threshold value, it is determined that the molding state to be detected has been reached.

[0056] (Notification unit) FIG. 9 shows an example of an alert message displayed on the mobile terminal T. When it is determined by the state estimation unit 30 that the molding state has reached a predetermined state based on the feature amount F extracted by the extraction unit 20, the notification unit 40 transmits an alert message together with the captured image to the mobile terminal T held by the supervisor who monitors the injection molding system 100.

[0057] The notification unit 40 associates with the accounts of SNS (Social Networking Service) such as "LINE" (trademark) and "X" (trademark) used by the supervisor, and transmits an alert message together with the captured image. For example, as shown in FIG. 9, on the LINE talk screen, a message indicating that the molding state has reached a predetermined state and the captured image are displayed. Note that the notification unit 40 may also display the captured image and the alert message on the display of a computer that executes the operation control and processing control of the imaging unit 10, the extraction unit 20, the state estimation unit 30, and the notification unit 40.

[0058] (3) Detection process of the monitoring device FIG. 10 is a flowchart showing the flow of the detection process for estimating the molding state of the injection molding system 100 in the monitoring device 1. Hereinafter, the detection process for estimating the molding state of the injection molding system 100 in the monitoring device 1 according to the present embodiment will be described with reference to the drawings.

[0059] In the monitoring device 1 according to the present embodiment, the imaging unit 10 is installed so as to face the mold split surface PL opened by the ejecting unit 124 of the ejecting device 120, and is configured to image the mold split surface PL including the cavity CA from the front.

[0060] First, in step S101, the imaging unit 10 is turned on, and the mold split surface PL of the unused mold K is imaged to obtain a reference image (S101). The captured image may be either a still image or a moving image. In the case of a moving image, it is obtained as a frame image extracted at regular time intervals.

[0061] Next, in step S102, with the first injection molding completed and the mold K opened, the mold split surface PL including the cavity CA is imaged from the front to obtain image data (S102). Then, in step S103, the feature amount F is extracted from the image data (S103). When the feature amount F is a black image when the obtained image data is binarized and the entire image is converted into a black-and-white image, first, the obtained image data is binarized. The binarization process sets the pixel value to "0" or "1" for each pixel in the image data with a predetermined threshold Th1 as the boundary. For example, when the luminance value is equal to or greater than the predetermined threshold Th1, the pixel value of the pixel is set to "1" (e.g., white), and when it is less than the threshold Th1, the pixel value of the pixel is set to "0" (e.g., black). Note that the feature amount F may be extracted as the contour line CT in the captured image, or may be a divided region RG divided into regions with a predetermined density and / or color specified in advance from the captured image.

[0062] In step S104, the difference between the extracted feature amount F and the feature amount F extracted from the reference image is calculated (S104). When the feature amount F is a black image when the entire image is converted into a black-and-white image by performing binarization processing, the ratio (black ratio BR) of the pixel values in the image data that has become a black-and-white two-tone image to the binarized reference image with "0" (black) is calculated (S104). Then, in step S105, it is determined whether the difference is greater than the threshold value, that is, whether the forming state has reached a predetermined state (S105). Specifically, it is determined whether the black ratio BR calculated in step S104 has reached a predetermined threshold value Th2, whether the extracted contour line CT has reached a predetermined level, and whether the extracted divided region RG has reached a predetermined level. When the difference is greater than the threshold value (S105; Yes), the notification unit 40 transmits an alert message together with the captured image to the mobile terminal T held by the supervisor (S106). When the difference has not reached the predetermined threshold value (S105; No), the process returns to step S102, and the mold parting surface PL including the cavity CA is imaged from the front in a state where the injection molding is completed and the mold K is opened, and the image data is acquired (S102).

[0063] (4) Function and effect of the monitoring device · In one molding cycle, the monitoring device 1 extracts the feature amount F from the image data obtained by imaging the mold parting surface PL including the cavity CA in the state where the mold is opened before closing the fixed-side mold K1 and the movable-side mold K2, and estimates the forming state based on the comparison with the reference image which is the first captured image data. Thereby, the image of the mold parting surface PL including the cavity CA can be acquired for each molding cycle, and the forming state can be monitored in real time.

[0064] · The monitoring device 1 performs binarization processing on the acquired image data using a preset threshold Th1 to convert the entire image into a black-and-white image, and estimates the forming state based on the change in the black image with respect to the reference image. Here, whether or not the forming state has reached a predetermined state is estimated based on whether or not the ratio of the black image (black ratio BR) has reached a predetermined value, or whether or not a pattern different from the reference image has been detected in the black image. Thereby, it is possible to reliably detect stains on the gas vent portion GV of the mold parting surface PL, burning of the mold parting surface PL, resin residue deposition on the mold parting surface PL including burrs, etc., and partial resin residue of the resin molded body S in the cavity CA.

[0065] · The monitoring device 1 may estimate the forming state based on the contour line CT extracted from the acquired image data. The extracted contour line CT indicates stains on the gas vent portion GV of the mold parting surface PL, etc., and it is possible to estimate the forming state by determining whether or not the contour line CT has reached a predetermined level.

[0066] · The monitoring device 1 can extract the divided region RG divided from the acquired image data into regions with a preset density and / or color, and estimate the forming state based on the change in the divided region RG with respect to the reference image.

[0067] · The monitoring device 1 can estimate the forming state from the feature amount F extracted by the extraction unit 20 using a learned model obtained by machine learning for estimating the forming state from the image data.

[0068] · The monitoring device 1 can calculate the similarity between the feature amount F extracted by the extraction unit 20 and a template image prepared in advance, and estimate the forming state based on the similarity.

[0069] · When it is determined that the forming state has reached a predetermined state, an alert message is transmitted to the portable terminal T of the monitor together with the captured image of the mold parting surface PL including the cavity CA. Thereby, the forming state can be intuitively and easily notified, and the monitoring timing of the injection molding machine 110 and the mold K can be reduced.

Description of Symbols

[0070] 1 ··· Monitoring device 10 ··· Imaging unit, 20 ··· Extraction unit, 30 ··· State estimation unit, 40 ··· Notification unit, 50 ··· Memory unit 100 ··· Injection molding system, 110 ··· Injection molding machine, 120 ··· Takeout device K ··· Mold, K1 ··· Fixed mold, K2 ··· Movable mold, GV ··· Gas vent part S ··· Resin molded body, T ··· Mobile terminal, F ··· Feature amount, CT ··· Contour line RG ··· Division area

Claims

1. A monitoring device for estimating a molding state during the molding process of a resin molded body, comprising: imaging means for imaging a mold parting surface capable of forming a mold cavity between a fixed mold and a movable mold opposing the fixed mold to obtain an imaging image; extracting means for extracting a feature amount of the imaging image; state estimation means for estimating the molding state based on the feature amount; notification means for notifying when the molding state reaches a predetermined state. The monitoring device is characterized by the above.

2. The imaging image is an image that is imaged each time before the fixed mold and the movable mold are closed in one molding cycle. The monitoring device according to claim 1, characterized by the above.

3. The imaging image is a frame image extracted at regular time intervals from a moving image of the mold parting surface. The monitoring device according to claim 2, characterized by the above.

4. The feature amount is a black image obtained by performing binarization processing using a threshold value preset in the imaging image to convert the entire image into a black-and-white image. The monitoring device according to claim 1, characterized by the above.

5. The feature amount is a contour line extracted from the imaging image. The monitoring device according to claim 1, characterized by the above.

6. The feature amount is a divided region divided into regions with a preset density and / or color from the imaging image. The monitoring device according to claim 1, characterized by the above.

7. The state estimation means estimates the molding state based on the difference between the feature amount and the feature amount extracted from a reference image which is the first imaged imaging image. The monitoring device according to any one of claims 1 to 6, characterized by the above.

8. The state estimation means estimates the molding state from the feature amount extracted by the extraction means using a learned model obtained by machine learning for estimating the molding state from the imaging image. The monitoring device according to any one of claims 1 to 6, characterized by the above.

9. The state estimation means calculates the similarity between the feature amount and a template image prepared in advance, and estimates the molding state based on the similarity. The monitoring device according to any one of claims 1 to 6, characterized by the above.

10. The imaging means is attached to the take-out device of the resin molded body, and images the mold parting surface when the fixed mold and the movable mold are open. The monitoring device according to claim 1, characterized in that...

Images