Mould processing injection molding method, device and equipment based on machine vision and medium

Abstract

The invention provides a mold processing injection molding method, device and equipment based on machine vision and a medium, and the method comprises the following steps: performing all-directional scanning on the surface of a mold cavity based on a machine vision system to obtain cavity three-dimensional image data of the mold cavity; inputting the three-dimensional image data of the cavity into an image analysis and detection model, and obtaining a first analysis and detection result based on a deep convolutional neural network and an image neural network; calling a spectrum analyzer to detect the to-be-injection-molded material based on the first analysis detection result to obtain a material characteristic detection result of the to-be-injection-molded material; determining material processing parameters of the to-be-injection-molded material according to the material characteristic detection result, and melting the to-be-injection-molded material based on the material processing parameters to obtain a material melt; and first mold injection molding machining parameters are determined based on the mold characteristic parameters of the mold cavity, and the material melt is injected into the mold cavity based on the first mold injection molding machining parameters. The product quality stability and the production efficiency are improved, and the production cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of mold processing, and in particular, to an injection molding method, device, equipment and medium for mold processing based on machine vision. Background Art

[0002] In the field of precision injection molding of molds, traditional processing methods usually rely on preset process parameters and the experience of operators to control the injection molding process. First, there may be certain fluctuations in the actual properties of materials between different batches, and it is difficult for preset process parameters to dynamically adapt to such changes. Therefore, it is easy to cause unstable quality of injection molded products, such as dimensional deviations, surface defects and other problems. Second, after long-term use, injection molds will show wear, deformation and other conditions. Traditional methods cannot timely and accurately sense these changes in mold states and adjust processing parameters accordingly, thus further affecting the accuracy and quality consistency of products. Moreover, for some molds with complex structures and high-precision injection molded products, traditional processing methods are difficult to precisely control the melt flow behavior during injection molding, which may cause defects such as uneven filling and air entrapment, reducing the yield rate of products and increasing production costs. In addition, traditional quality inspection mainly relies on sampling inspection, which is not only inefficient, but also cannot monitor the production process of each product in real time, and cannot timely detect and correct abnormal conditions in the production process. Once batch quality problems occur, it will cause relatively large economic losses. Summary of the Invention

[0003] The present invention provides an injection molding method, device, equipment and medium for mold processing based on machine vision, which effectively overcomes the defects of traditional processing methods through comprehensive and precise monitoring of mold states, material properties, injection molding processes and product quality, improves the quality stability of products, production efficiency and reduces production costs.

[0004] In a first aspect, the present invention provides an injection molding method for mold processing based on machine vision, including:

[0005] Performing an omnidirectional scan on the surface of a mold cavity based on a machine vision system to obtain three-dimensional image data of the mold cavity;

[0006] Inputting the three-dimensional image data of the mold cavity into the image analysis and detection model, and performing feature extraction on the three-dimensional image data of the mold cavity based on a deep convolutional neural network layer in the image analysis and detection model to obtain local cavity features;

[0007] Performing feature extraction on the three-dimensional image data of the mold cavity based on a graph neural network layer in the image analysis and detection model to obtain topological structure features of the mold cavity;

[0008] Based on the fully connected layer in the image analysis detection model, connect the local cavity features and the cavity topological structure features to obtain the first analysis and detection result output by the image analysis detection model;

[0009] Based on the first analysis and detection result, call a spectroscopic analyzer to detect the material to be injection-molded, and obtain the material property detection result of the material to be injection-molded;

[0010] Determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0011] Based on the mold property parameters of the mold cavity, determine the first mold injection processing parameters, and inject the material melt into the mold cavity based on the first mold injection processing parameters.

[0012] In a second aspect, the present invention provides a mold processing and injection device based on machine vision, which is applied to the mold processing and injection method based on machine vision as described in the first aspect; the mold processing and injection device based on machine vision includes:

[0013] An image acquisition module, configured to perform an omnidirectional scan on the surface of the mold cavity based on a machine vision system to obtain the three-dimensional image data of the cavity of the mold cavity;

[0014] A model detection module, configured to:

[0015] Input the three-dimensional image data of the cavity into the image analysis detection model, and perform feature extraction on the three-dimensional image data of the cavity based on the deep convolutional neural network layer in the image analysis detection model to obtain local cavity features;

[0016] Perform feature extraction on the three-dimensional image data of the cavity based on the graph neural network layer in the image analysis detection model to obtain cavity topological structure features;

[0017] Based on the fully connected layer in the image analysis detection model, connect the local cavity features and the cavity topological structure features to obtain the first analysis and detection result output by the image analysis detection model;

[0018] A material detection module, configured to detect the material to be injection-molded by calling a spectroscopic analyzer based on the first analysis and detection result, and obtain the material property detection result of the material to be injection-molded;

[0019] A processing module, configured to determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0020] The injection molding module is used to determine the first mold injection molding processing parameters based on the mold characteristic parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection molding processing parameters.

[0021] In a third aspect, the present invention further provides an electronic device, including: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing the machine vision-based mold processing and injection molding method as described in any one of the above.

[0022] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, in which a computer software program is stored, and when the computer software program is executed by a processor, the machine vision-based mold processing and injection molding method as described in any one of the above is implemented.

[0023] In a fifth aspect, the present invention further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the machine vision-based mold processing and injection molding method as described in any one of the above is implemented.

[0024] In the machine vision-based mold processing and injection molding method provided by the embodiments of the present invention, in the mold state detection step, through the precise detection of the mold cavity by the machine vision system, potential problems of the mold can be discovered and processed in time, avoiding product quality problems caused by mold defects and ensuring the stability of the product appearance quality. The material characteristic analysis step ensures that the injection molding parameters can be precisely adjusted according to the actual characteristics of the material, effectively compensating for the differences between different batches of materials, making the performance such as the fluidity and formability of the melt more stable, reducing product size deviations and internal defects caused by material fluctuations, and improving the product quality. Since the mold state detection and material characteristic analysis can be quickly completed before injection molding, and the injection molding parameters can be automatically optimized and adjusted, the number of trial moldings and adjustment time caused by mold problems and material differences are reduced, the initial success rate of injection molding production is improved, the production cycle of the product is shortened, and the production efficiency is increased. The improvement of product quality stability reduces the generation rate of waste products and defective products, reduces the waste of raw materials, and directly reduces the production cost. Description of the Drawings

[0025] Figure 1 is a flowchart of the machine vision-based mold processing and injection molding method provided by the embodiments of the present invention;

[0026] Figure 2 is a structural diagram of the machine vision-based mold processing and injection molding device provided by the embodiments of the present invention;

[0027] Figure 3 is an embodiment diagram of the electronic device provided by the embodiments of the present invention;

[0028] Figure 4 It is an embodiment diagram of the computer-readable storage medium provided by the embodiments of the present invention. Detailed implementation manners

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0030] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0031] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or more advantageous than other embodiments. The following description is given to enable any person skilled in the art to make and use the present invention. In the following description, details are set forth for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without these specific details. In other instances, well-known structures and processes are not described in detail to avoid unnecessary details from obscuring the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0032] Refer to Figure 1 as shown Figure 1 is a flowchart of the mold processing and injection molding method based on machine vision provided by the present invention. In the embodiments of the present invention, the execution subject of the mold processing and injection molding method based on machine vision is an injection molding processing device. Therefore, the mold processing and injection molding method based on machine vision includes:

[0033] Step 10, perform an all-round scan on the surface of the mold cavity based on the machine vision system to obtain the three-dimensional image data of the mold cavity.

[0034] Optionally, a machine vision system is installed in the injection molding device according to an embodiment of the present invention. The machine vision system includes a plurality of laser emitters and a high-resolution CCD camera. Therefore, the injection molding device can call the plurality of laser emitters and the high-resolution CCD camera to perform an omnidirectional scan on the surface of the mold cavity, and obtain the three-dimensional image data of the mold cavity, as specifically described in steps 101 to 103.

[0035] Step 20: Input the three-dimensional image data of the cavity into the image analysis and detection model, and perform feature extraction on the three-dimensional image data of the cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain the local features of the cavity.

[0036] Optionally, an image analysis and detection model is embedded in the injection molding device according to an embodiment of the present invention. The image analysis and detection model is trained based on the sample image data and its corresponding analysis and detection result labels.

[0037] Optionally, an architecture combining a deep convolutional neural network (DCNN) and a graph neural network (GNN) is adopted in an embodiment of the present invention. The three-dimensional image data of the cavity is input into the image analysis and detection model, and the deep convolutional neural network layer DCNN in the image analysis and detection model performs feature extraction on the three-dimensional image data of the cavity to obtain the local features of the cavity in the three-dimensional image data of the cavity, such as surface texture, micro defects, etc. In one embodiment, the convolution kernel size of the l-th layer in the deep convolutional neural network layer DCNN is k l , the stride is s l , the number of input channels is c l , the number of output channels is d l , then the size O l of the output feature map of this layer = (I l - k l + 2p l ) / (s l ) + 1, where I l is the size of the input feature map, p l is the padding size, and each element of the feature map is:

[0038]

[0039] where K l is the convolution kernel of the l-th layer, and b l is the bias term of the l-th layer.

[0040] Step 30: Perform feature extraction on the three-dimensional image data of the cavity based on the graph neural network layer in the image analysis and detection model to obtain the topological structure features of the cavity.

[0041] Further, based on the graph neural network layer GNN in the image analysis detection model, feature extraction is performed on the three-dimensional image data of the cavity, such as analyzing the overall topological structure of the cavity to obtain the cavity topological structure features, for example, detecting the connectivity of the mold cavity, complex contour features, etc. In one embodiment, the surface of the mold cavity is discretized into N nodes V = {v1, v2,..., v N}, the edge set is E = {(v i , v j )|i, j ∈ {1, 2,..., N}}, and the information transfer formula of GNN is:

[0042]

[0043] Among them, is the feature vector of node v i at the t-th iteration, N(i) is the set of neighbor nodes of node v i , W edge and W self are learnable weight matrices, σ() is an activation function (such as the ReLU function), and b g is a bias vector.

[0044] Among them, in the training process of the image analysis detection model in the embodiment of the present invention, the backpropagation algorithm is adopted, and the loss function is: Among them, M is the number of training samples, y k is the true label, is the model prediction result, μ is the regularization coefficient, W l (i, j) is the weight element of the l-th layer of the model, and by continuously adjusting the network parameters, the loss function is minimized.

[0045] Step 40, based on the fully connected layer in the image analysis detection model, the local features of the cavity and the cavity topological structure features are connected to obtain the first analysis and detection result output by the image analysis detection model.

[0046] Further, based on the fully connected layer in the image analysis detection model, the local features of the cavity and the cavity topological structure features are connected to obtain the first analysis and detection result. Among them, the first analysis and detection result in the embodiment of the present invention includes the wear degree D w of the mold cavity, the surface roughness R a , whether there is foreign object attachment F a , etc. For example, the wear degree D w can be obtained by calculating the volume loss of the key parts of the cavity. In one embodiment, the original volume of the key part is V0, and the current volume is V1, then D w = (V0 - V1) / (V0)*100%, and the predicted value of the wear degree output by the model It is obtained by calculation of the fully connected layer based on the features extracted by the above DCNN and GNN:

[0047] Among them, f dcnn and f gnn are the wear-related feature vectors extracted by DCNN and GNN respectively, ω1 and ω2 are weight parameters, and b d is the bias term.

[0048] Step 50, based on the first analysis and detection result, call a spectral analyzer to detect the material to be injection-molded, and obtain the material property detection result of the material to be injection-molded.

[0049] Furthermore, the injection molding device identifies the first analysis and detection result to determine whether the first analysis and detection result meets the preset result. Among them, the preset result is that the wear degree in the first analysis and detection result is greater than the preset wear degree, the surface roughness in the first analysis and detection result is greater than the preset roughness, etc.

[0050] Furthermore, if it is determined that the first analysis and detection result meets the preset result, the injection molding device then calls a spectral analyzer to detect the material to be injection-molded, and obtains the material property detection result of the material to be injection-molded, specifically as described in steps 501 to 503.

[0051] Step 60, determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt.

[0052] Optionally, the material processing parameters in the embodiments of the present invention include melting temperature T m , screw speed n s , back pressure P b , etc. Taking the melting temperature T m as an example, it is determined according to the crystallinity X c of the material, thermal conductivity k and specific heat capacity c. In one embodiment, the ideal melting temperature range of the material is [T min , T max , and a fuzzy logic control algorithm is used to determine it. The fuzzy set of crystallinity X c is {low, medium, high}, the fuzzy set of thermal conductivity k is {small, medium, large}, the fuzzy set of specific heat capacity c is {low, medium, high}, and the fuzzy set of melting temperature T m is {low temperature, medium temperature, high temperature}. Establish a fuzzy rule table. For example: if X c is high and k is small and c is low, then T m is high temperature; if X c is medium and k is medium and c is medium, then T m is medium temperature, etc. Reason according to the fuzzy rules. In one embodiment, μ k (y), μ c (z) are the membership function values of crystallinity, thermal conductivity, and specific heat capacity in their respective fuzzy sets. For the i-th fuzzy rule R i : If X c is A i and k is B i and c is C i , then T m is D i , and its activation strength

[0053] The final melting temperature T m The crisp value of is calculated by the centroid method:

[0054] where is the central value of the melting temperature fuzzy set corresponding to the i-th fuzzy rule.

[0055] Optionally, the determination of the screw speed n s and the back pressure P b also adopts a similar algorithm considering multiple factors. For example, a complex functional relationship or fuzzy rule model is established based on factors such as the melt index MI of the material and the density ρ of the material to determine. During the melting process, the actual melting state of the material is monitored in real time through temperature sensors, pressure sensors, etc., and the PID control algorithm is used to adjust the heating power P h and the speed n m of the screw drive motor to ensure that the material can be stably melted to the target state. The PID control algorithm formula is:

[0056] where μ(t) is the control output (such as the adjustment amount of heating power or motor speed), e(t) is the deviation between the actual value and the target value (such as the deviation between the actual temperature and the melting temperature T m ), and K p , K i , K d are the proportional, integral, and differential coefficients respectively, and these coefficients can be adjusted through experiments or adaptive algorithms according to the material characteristics and equipment characteristics.

[0057] Furthermore, the injection molding processing device melts the material to be injection-molded through the material processing parameters to obtain the material melt of the material to be injection-molded.

[0058] Step 70, determine the first mold injection molding processing parameters based on the mold characteristic parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection molding processing parameters.

[0059] Further, the injection molding device determines the first mold injection molding parameters according to the mold characteristic parameters of the mold cavity, as specifically described in steps 701 to 705, and injects the material melt into the mold cavity according to the first mold injection molding parameters.

[0060] In the embodiment of the present invention, through the precise detection of the mold cavity by the machine vision system, potential problems of the mold can be detected and processed in a timely manner, avoiding product quality problems caused by mold defects and ensuring the stability of the product appearance quality. On the other hand, it ensures that the injection molding parameters can be precisely adjusted according to the actual characteristics of the material, effectively compensating for the differences between different batches of materials, making the properties such as the fluidity and formability of the melt more stable, reducing product size deviations and internal defects caused by material fluctuations, and improving product quality. Since the mold state detection and material characteristic analysis can be quickly completed before injection molding, and the injection molding parameters can be automatically optimized and adjusted, the number of trial moldings and adjustment time caused by mold problems and material differences are reduced, the initial success rate of injection molding production is improved, the production cycle of the product is shortened, and the production efficiency is increased. The improvement of product quality stability reduces the generation rate of waste products and defective products, reduces the waste of raw materials, and directly reduces the production cost.

[0061] In one embodiment, the descriptions of steps 101 to 103 are as follows:

[0062] Step 101, obtain the angle θ between the laser emitter and the surface of the mold cavity, the angle α between the high-resolution CCD camera and the laser beam emitted by the laser emitter, the displacement Δx of the laser spot formed on the surface of the mold cavity on the imaging plane of the high-resolution CCD camera, and the focal length f of the high-resolution CCD camera;

[0063] Step 102, based on the angle θ, the angle α, the displacement Δx, and the focal length f, determine the height h of any point on the surface of the mold cavity. The calculation formula for the height h is h = (Δx * f) / (tanθ * tanα);

[0064] Step 103, process and splice the images of the surface of the mold cavity collected by the high-resolution CCD camera at multiple viewing angles, calculate the surface height information of the mold cavity corresponding to each pixel point, and construct the three-dimensional image data of the mold cavity.

[0065] Specifically, the embodiment of the present invention adopts the multi-view laser triangulation method, that is, multiple laser emitters and corresponding high-resolution CCD cameras are arranged around the mold cavity. The laser emitter emits a laser beam at a specific angle to the surface of the mold cavity, and the laser forms a light spot on the surface of the mold cavity. Due to the height change of the surface of the mold cavity, the angle of the reflected light will also change. The high-resolution CCD camera captures images of the reflected light from different viewing angles.

[0066] In one embodiment, the included angle between the laser emitter and the surface of the mold cavity is θ, the included angle between the high-resolution CCD camera and the laser beam is α, the displacement of the laser spot on the imaging plane of the high-resolution CCD camera is Δx, and the focal length of the high-resolution CCD camera is f. According to the principle of similar triangles, the calculation formula for the height h of a certain point on the surface of the mold cavity can be derived: h = (Δx * f) / (tanθ * tanα).

[0067] By processing and stitching the images from multiple perspectives and combining the above formula, the surface height information of the mold cavity corresponding to each pixel point is calculated, thereby constructing the three-dimensional image data of the cavity. For example, for the image at the i-th perspective, the height corresponding to its pixel coordinates (x i , y i ) is h i (x i , y i ). After processing n perspectives, the final three-dimensional image data of the cavity where ω i is the weight coefficient of the i-th perspective and can be dynamically adjusted according to the importance and accuracy of the perspective. For example β i is the included angle between the i-th perspective and the central axis of the cavity.

[0068] In the embodiment of the present invention, the image analysis and detection model can accurately obtain the analysis and detection results. Therefore, subsequent precision injection molding of the mold can be carried out according to the analysis and detection results, improving the quality stability of the product.

[0069] In one embodiment, the descriptions of steps 501 to 503 are as follows:

[0070] Step 501, if it is determined based on the first analysis and detection result that the surface of the mold cavity meets the preset result, then a spectral analyzer is called to emit multiple optical bands to detect the material to be injection-molded, and the reflectivity, absorptivity, and transmittance of the material to be injection-molded for each optical band emitted by the spectral analyzer are obtained;

[0071] Step 502, based on the reflectivity, absorptivity, and transmittance of the material to be injection-molded for each optical band emitted by the spectral analyzer, a relationship model between the material properties of the material to be injection-molded and the optical bands of the spectral analyzer is established;

[0072] Step 503, analyze the relationship model to obtain the detection result of the material properties of the material to be injection-molded.

[0073] Optionally, if it is determined based on the first analysis and detection result that the surface of the mold cavity does not meet the preset result, where the preset result is that the wear degree in the first analysis and detection result is greater than the preset wear degree, and the surface roughness in the first analysis and detection result is greater than the preset roughness, adjust the detection parameters of the spectral analyzer according to the first analysis and detection result of the mold cavity. For example, if the wear degree of the mold cavity is high or the surface roughness is large, it may affect the flow and molding of the material in the cavity. At this time, the spectral analyzer will increase the detection accuracy of the viscosity-related characteristics of the material. In one embodiment, the wear degree threshold of the mold cavity is T w , when D w >T w , the detection accuracy of the viscosity coefficient of the material by the spectral analyzer is increased by k times (k>1).

[0074] If it is determined based on the first analysis and detection result that the surface of the mold cavity meets the preset result, call the spectral analyzer to emit multiple light bands to detect the material to be injection-molded. In the embodiment of the present invention, the spectral analyzer adopts multi-band spectral analysis technology, emits light in multiple bands from ultraviolet to infrared to irradiate the material to be injection-molded, and detects the absorption, reflection, and transmission characteristics of the material for light in different bands, and obtains the reflectivity, absorptivity, and transmittance of the material to be injection-molded for each light band emitted by the spectral analyzer. In one embodiment, the light bands emitted by the spectral analyzer are λ={λ1, λ2,.., λ P}; the reflectivity of the material for the p-th band of light is r p , the absorptivity is a p , the transmittance is t p , and r p +a p +t p =1.

[0075] Further, based on the reflectivity, absorptivity, and transmittance of the material to be injection-molded for each light band emitted by the spectral analyzer, establish a relationship model between the material characteristics of the material to be injection-molded and the light bands of the spectral analyzer, and analyze the relationship model to obtain the material characteristics detection result of the material to be injection-molded, where the material characteristics detection result includes material composition information, material purity information, and material physical property parameters. In one embodiment, for the crystallinity X c of the material, an algorithm combining multiple linear regression and neural network is adopted:

[0076]

[0077] where β p is the linear regression coefficient; N(·) is a small neural network for capturing the non-linear relationship between spectral data; γ is the weight coefficient of the neural network, b sis the bias term. The structure of the neural network can be a simple multi-layer perceptron. For example, the input layer has P nodes (corresponding to the reflectance of P bands), the hidden layer has Q nodes, and the output layer has 1 node (corresponding to the predicted value of crystallinity). The activation function of the hidden layer can adopt the Sigmoid function σ(x) = 1 / (1 + e -x ), and the calculation formula of the output layer is:

[0078]

[0079] where, v pq , ω q , ω are the weight parameters of the neural network, and b q , b y are the bias terms.

[0080] The embodiments of the present invention ensure that the injection molding parameters can be precisely adjusted according to the actual characteristics of the material, effectively compensate for the differences between different batches of materials, make the properties such as the fluidity and formability of the melt more stable, reduce the product size deviation and internal defects caused by material fluctuations, and improve the product quality.

[0081] In one embodiment, the descriptions of steps 701 to 705 are as follows:

[0082] Step 701, obtain the temperature change range of the mold cavity during the injection molding process and the heat exchange rate between the mold cavity and the surrounding environment; Step 702, calculate the heat balance time based on the temperature change range, heat exchange rate and heat capacity, and determine the injection molding cycle of the mold cavity based on the heat balance time; Step 703, determine the injection molding speed of the mold cavity based on the filling volume and injection molding cycle; Step 704, determine the injection molding pressure of the mold cavity based on the minimum wall thickness and injection molding speed; Step 705, determine the injection molding cycle, injection molding speed and injection molding pressure as the first mold injection molding processing parameters.

[0083] Optionally, the mold characteristic parameters include the heat capacity C m of the mold, the cooling water channel structure parameters of the mold (such as the water channel diameter d w , the water channel length L w , the water channel spacing s w , etc.), the material thermal conductivity k m of the mold, etc.

[0084] First, calculate the heat balance time t b of the mold, and adopt the simplified model of the finite element thermal analysis algorithm: t b = (C m *ΔT) / q, where ΔT is the temperature change range of the mold during the injection molding process, and q is the heat exchange rate between the mold and the surrounding environment:

[0085] q = ∑ i (2πk m L i )*(T i - T env ) / [ln(d i+1 / d i )] (where L i is the length of different parts of the mold, d i+1 , d i are the diameters or thicknesses of adjacent two-layer mold structures, T i is the temperature of the corresponding part, and T env is the ambient temperature. According to the thermal equilibrium time t b of the mold, the injection molding cycle T c is determined. For example, T c = t b + Δt, where Δt is the safety margin time.)

[0086] Further, the determination of the injection molding pressure P i takes into account the complexity C c of the mold cavity (which can be quantified by factors such as the surface area to volume ratio of the cavity and the number of parting surfaces), the viscosity μ of the material melt, and the injection molding speed v i . An empirical formula combined with numerical simulation is used:

[0087] where L c is the flow path length of the melt in the cavity, h c is the minimum wall thickness of the cavity, P0 is the basic injection molding pressure, is the empirical coefficient (determined according to different materials and mold types).

[0088] Further, the injection molding speed v i is determined according to the filling volume V f of the mold cavity, the flow rate Q of the material melt, and the injection molding time t i (t i is the injection molding stage time in the injection molding cycle T c ): v i = (Q * t i ) / V f , where the flow rate Q of the material melt can be calculated according to the screw diameter D s , the screw speed n s and the compression ratio r c of the melt:

[0089] Further, the holding pressure P h and the holding time t h are based on the shrinkage rate S r of the material and the volume V of the mold cavityc Determined by the pressure - volume - temperature (PVT) characteristics of the material melt. Numerical calculation method based on PVT data: First, establish a PVT model V = f(P, T) according to the PVT data of the material, and then during the holding pressure stage, calculate the holding pressure and time according to the shrinkage of the material. For example, in one embodiment, the volume change of the material during the holding pressure stage is ΔV, then the holding pressure P h =(ΔV * k PVT ) / V c (k PVT is the pressure - volume coefficient calculated according to the PVT model), and the holding time t h is determined according to the solidification time t s of the material and the decay characteristics of the holding pressure. For example, t h =β * t s (β is the decay coefficient, determined according to the material and mold characteristics).

[0090] Furthermore, during the injection molding process, parameters such as pressure and melt flow position during the injection molding process are monitored in real - time through a pressure sensor, displacement sensor, etc., and the injection molding parameters are dynamically adjusted using an adaptive control algorithm. For example, if it is monitored that the injection molding pressure exceeds the preset upper limit P max , then the injection molding speed v i is appropriately reduced, and the adjustment formula is At the same time, the injection molding cycle and holding pressure parameters are recalculated according to the new injection molding speed to ensure the stability of the injection molding process and the consistency of product quality.

[0091] In the embodiment of the present invention, since the mold state detection and material property analysis can be quickly completed before injection molding, and the injection molding parameters can be automatically optimized and adjusted, the number of trial moldings and adjustment time caused by mold problems and material differences are reduced, the initial success rate of injection molding production is improved, the production cycle of the product is shortened, and the production efficiency is increased. The improvement of product quality stability reduces the generation rate of defective and substandard products, reduces the waste of raw materials, and directly reduces the production cost.

[0092] In one embodiment, the mold processing and injection molding method based on machine vision further includes:

[0093] Step a1, during the process of injecting the material melt into the mold cavity with the first mold injection molding parameters, monitor the flow state information of the material melt in the mold cavity; Step a2, adjust the first mold injection molding parameters based on the flow state information to obtain the second mold injection molding parameters; Step a3, inject the material melt into the mold cavity based on the second mold injection molding parameters.

[0094] Optionally, the flow state information in the embodiments of the present invention includes the melt front position, melt flow velocity, and melt filling uniformity. Therefore, during the process of injecting the material melt into the mold cavity with the injection molding parameters of the first mold, the melt front position, melt flow velocity, and melt filling uniformity of the material melt in the mold cavity are monitored.

[0095] In one embodiment, the monitoring of the melt front position is specifically as follows:

[0096] Use a machine vision system to take real-time pictures of the inside of the mold cavity. A plurality of marking points are preset inside the mold cavity (these marking points can be tiny reflective or fluorescent material points, which do not affect injection molding). The melt front position is determined by identifying the relative positions of the marking points and the melt front in the image.

[0097] In one embodiment, the resolution of the image taken by the machine vision system is M*N pixels, and the coordinates of the marking point in the image are (x m , y m ). The coordinates of multiple feature points (x fp , y fp ) on the melt front contour are determined through an image recognition algorithm, where p = 1, 2,..., P.

[0098] Calculate the distance d from the melt front to the marking point fp , and adopt a calculation method based on pixel distance:

[0099] where k is the proportional coefficient between the image pixels and the actual size, which is obtained through prior calibration.

[0100] Then, based on the distance information of multiple feature points, determine the overall position distribution of the melt front. For example, by calculating the average value and standard deviation of these distances to evaluate the concentration degree and dispersion degree of the melt front position.

[0101] In one embodiment, the monitoring of the melt flow velocity is specifically as follows:

[0102] Adopt the time series image analysis method. Take multiple frames of images of the inside of the mold cavity at continuous time intervals Δt, and compare the changes in the melt front position in different frames of images.

[0103] In one embodiment, the coordinates of the melt front feature points at the t1 moment are (x fp1 , y fp1 ), and the corresponding coordinates at the t2 = t1 + Δt moment are (x fp2 , y fp2 ).

[0104] Then the flow velocity of the melt in the direction of the characteristic point

[0105] For the entire melt front, after calculating the flow velocities of multiple characteristic points, the average flow velocity of the melt is determined by the method of weighted average where w p is the weight coefficient of the characteristic point, which can be determined according to the importance of the position of the characteristic point to the overall flow. For example, the characteristic points near the gate position have a larger weight, and w p can be set to d min / (d min +d fp )(d min is a very small value to prevent the denominator from being zero).

[0106] In one embodiment, the monitoring of the melt filling uniformity is as follows:

[0107] The mold cavity is divided into multiple sub-regions A i , i = 1, 2,.., I, and the filling height h of the melt in each sub-region is analyzed through a machine vision system i . In one embodiment, there is a functional relationship h i = f(g i )(where g i is the average gray value of the melt filling region in the sub-region A i ), and this functional relationship is obtained through pre-experimental calibration, such as polynomial fitting (α j is the fitting coefficient). Calculate the filling uniformity index where The smaller the U value, the more uniform the filling

[0108] Furthermore, according to the flow state information, the first mold injection processing parameters are adjusted to obtain the second mold injection processing parameters, where the second mold injection processing parameters include injection pressure adjustment, injection speed adjustment, and holding pressure parameter adjustment

[0109] In one embodiment, the specific process of injection pressure adjustment is as follows:

[0110] When the melt front position deviates far from the expected position (such as the average distance is greater than the preset threshold D th ) or the flow velocity is uneven (such as the velocity standard deviation σ d is greater than V th ), the injection pressure needs to be adjusted

[0111] In one embodiment, the initial injection pressure is P i1 , and the adjusted injection pressure Pi2 Adopt an adjustment algorithm based on fuzzy logic. Define the deviation of the melt front position (D ideal is the distance corresponding to the ideal melt front position) of the fuzzy set as {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, the flow velocity deviation ΔV = σ d - V ideal (V ideal is the standard deviation of the ideal flow velocity) of the fuzzy set as {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, and the fuzzy set of the injection pressure adjustment amount ΔP is {substantially reduced, moderately reduced, slightly reduced, unchanged, slightly increased, moderately increased, substantially increased}.

[0112] Establish a fuzzy rule table. For example: If ΔD is negative large and ΔV is negative large, then ΔP is substantially increased; if ΔD is positive small and ΔV is zero, then ΔP is slightly increased, etc.

[0113] Inference is carried out according to the fuzzy rules. In one embodiment, μ ΔD (x), μ ΔV (y) are respectively the membership function values of the melt front position deviation and the flow velocity deviation in their respective fuzzy sets. For the kth fuzzy rule R k : If ΔD is A k and ΔV is B k , then ΔP is C k , and its activation strength α k = min(μ ΔD (x), μ ΔV (y)).

[0114] The clear value of the final injection pressure adjustment amount is calculated by the centroid method: ΔP = (∑ k α k *ΔP k ) / (∑ k α k ), where ΔP k is the central value of the fuzzy set of the injection pressure adjustment amount corresponding to the P k th fuzzy rule.

[0115] Then the adjusted injection pressure P i2 = P i1 +ΔP.

[0116] In one embodiment, the specific process for adjusting the injection speed is as follows:

[0117] When the melt filling uniformity is poor (such as U is greater than U th ), adjust the injection speed. In one embodiment, the initial injection speed is v i1 , and the adjusted injection speed is v i2 .

[0118] Adopt an adjustment algorithm based on a neural network. The melt filling uniformity index U and the thermal balance state parameter T of the mold cavity b (such as the temperature difference at each part of the cavity), the current viscosity μ of the material melt c are used as the inputs of the neural network. The neural network structure is a multi-layer perceptron. The input layer has 3 nodes, the hidden layer H has [number of nodes] nodes, and the output layer has 1 node (corresponding to the injection speed adjustment ratio β). The activation function of the hidden layer uses the hyperbolic tangent function tanh(x)=(e x -e -x ) / (e x +e -x ). The calculation formula of the output layer is:

[0119] where x j is the input parameters U, T b , μ c , ω, w h , v jh are the weight parameters of the neural network, b h , b y are the bias terms).

[0120] Then the adjusted injection speed v i2 =v i1 *(1 + β).

[0121] In one embodiment, the specific process for adjusting the holding pressure parameters is as follows:

[0122] If it is found during the injection process that the melt shrinkage does not match the expectation (such as by monitoring the internal pressure change of the mold cavity and the melt volume change), adjust the holding pressure and the holding time.

[0123] In one embodiment, the initial holding pressure is P h1 , the initial holding time is t h1 , the adjusted holding pressure P h2 and the holding time t h2 .

[0124] First, calculate the actual shrinkage rate S of the melt ra =(V m1 -V m2 ) / V m1 (where V m1 is the theoretical volume of the melt at the end of injection, and V m2 is the actually measured melt volume).

[0125] The holding pressure adjustment adopts a proportional control algorithm: P h2 =P h1*(1 + S ra ).

[0126] The holding pressure time adjustment adopts an algorithm based on the exponential function: t h2 = t h1 * exp(δ * S ra ), where δ is the exponential coefficient, which is determined according to the curing characteristics of the material and the thermal characteristics of the mold.

[0127] Furthermore, the material melt is injected into the mold cavity according to the injection molding processing parameters of the second mold, that is, according to the adjusted injection molding processing parameters of the second mold (including P i2 , v i2 , P h2 , t h2 , etc.), and the injection molding operation is continued. During the injection molding process, the flow state information of the melt is continuously monitored, and the above steps of adjusting the parameters are repeated until the injection molding process is completed, ensuring that the flow state of the material melt in the mold cavity is always in an ideal state during the entire injection molding process, thereby improving the quality and consistency of the injection molded product.

[0128] In one embodiment, the mold processing and injection molding method based on machine vision further includes: after the mold injection molding is completed and before the product is demolded, the mold injection molded product is comprehensively detected by the machine vision system to obtain the three-dimensional image data of the product of the mold injection molded product, and the three-dimensional image data of the product is input into the image analysis and detection model to obtain the second analysis and detection result output by the image analysis and detection model. The specific process is as in steps 10 to 40, which will not be elaborated here. Furthermore, if it is determined that the second analysis and detection result meets the preset result, the mold injection molded product is demolded.

[0129] In the embodiment of the present invention, the product is comprehensively detected before demolding, which can timely detect and eliminate unqualified products, avoid mixing defective products into qualified products, and ensure the stability and reliability of the product quality.

[0130] Furthermore, the mold processing and injection molding system based on machine vision provided by the present invention is described below. The mold processing and injection molding system based on machine vision described below can be mutually corresponded and referred to the mold processing and injection molding method described above.

[0131] Optionally, referring to Figure 2 , Figure 2 is the structural diagram of the mold processing and injection molding device based on machine vision provided by the present invention. The mold processing and injection molding device based on machine vision includes:

[0132] An image acquisition module 210, which is used to perform an omnidirectional scan on the surface of the mold cavity based on the machine vision system to obtain the three-dimensional image data of the mold cavity;

[0133] The model detection module 220 is used for:

[0134] Input the three-dimensional image data of the cavity into the image analysis and detection model, and extract features from the three-dimensional image data of the cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain the local features of the cavity;

[0135] Extract features from the three-dimensional image data of the cavity based on the graph neural network layer in the image analysis and detection model to obtain the topological structure features of the cavity;

[0136] Connect the local features of the cavity and the topological structure features of the cavity based on the fully connected layer in the image analysis and detection model to obtain the first analysis and detection result output by the image analysis and detection model;

[0137] The material detection module 230 is used to detect the material to be injection-molded by calling a spectroscopic analyzer based on the first analysis and detection result to obtain the material property detection result of the material to be injection-molded;

[0138] The processing module 240 is used to determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0139] The injection molding module 250 is used to determine the first mold injection molding parameters based on the mold property parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection molding parameters.

[0140] In the mold state detection step of the embodiment of the present invention, through the precise detection of the mold cavity by the machine vision system, potential problems of the mold can be discovered and processed in time, avoiding product quality problems caused by mold defects, and ensuring the stability of the product appearance quality. The material property analysis step ensures that the injection molding parameters can be precisely adjusted according to the actual properties of the material, effectively compensating for the differences between different batches of materials, making the properties such as the fluidity and formability of the melt more stable, reducing product size deviations and internal defects caused by material fluctuations, and improving product quality. Since the mold state detection and material property analysis can be quickly completed before injection molding, and the injection molding parameters can be automatically optimized and adjusted, the number of trial moldings and adjustment time caused by mold problems and material differences are reduced, the initial success rate of injection molding production is improved, the production cycle of the product is shortened, and the production efficiency is increased. The improvement of product quality stability reduces the generation rate of waste products and defective products, reduces the waste of raw materials, and directly reduces the production cost.

[0141] Please refer to Figure 3 , Figure 3 which is the embodiment diagram of the electronic device provided by the embodiment of the present invention. As Figure 3As shown in the figure, an embodiment of the present invention provides an electronic device 300, which includes a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, the following steps are implemented:

[0142] Perform an omnidirectional scan of the surface of the mold cavity based on a machine vision system to obtain three-dimensional image data of the mold cavity;

[0143] Input the three-dimensional image data of the mold cavity into an image analysis and detection model, and perform feature extraction on the three-dimensional image data of the mold cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain local features of the mold cavity;

[0144] Perform feature extraction on the three-dimensional image data of the mold cavity based on the graph neural network layer in the image analysis and detection model to obtain topological structure features of the mold cavity;

[0145] Connect the local features of the mold cavity and the topological structure features of the mold cavity based on the fully connected layer in the image analysis and detection model to obtain a first analysis and detection result output by the image analysis and detection model;

[0146] Call a spectral analyzer to detect the material to be injection-molded based on the first analysis and detection result to obtain a material property detection result of the material to be injection-molded;

[0147] Determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0148] Determine first mold injection molding processing parameters based on the mold characteristic parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection molding processing parameters.

[0149] Please refer to Figure 4 , Figure 4 is an embodiment diagram of a computer-readable storage medium provided by an embodiment of the present invention. As Figure 4 shown, this embodiment provides a computer-readable storage medium 400, on which a computer program 311 is stored. When the computer program 311 is executed by a processor, the following steps are implemented:

[0150] Perform an omnidirectional scan of the surface of the mold cavity based on a machine vision system to obtain three-dimensional image data of the mold cavity;

[0151] Input the three-dimensional image data of the mold cavity into an image analysis and detection model, and perform feature extraction on the three-dimensional image data of the mold cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain local features of the mold cavity;

[0152] Extract the features of the three-dimensional image data of the cavity using the graph neural network layer in the image analysis detection model to obtain the cavity topological structure features;

[0153] Connect the local features of the cavity and the cavity topological structure features using the fully connected layer in the image analysis detection model to obtain the first analysis and detection result output by the image analysis detection model;

[0154] Based on the first analysis and detection result, call a spectral analyzer to detect the material to be injection-molded to obtain the material property detection result of the material to be injection-molded;

[0155] Determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0156] Determine the first mold injection processing parameters based on the mold property parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection processing parameters.

[0157] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the mold processing and injection method based on machine vision provided by the above-mentioned various methods. The method includes:

[0158] Perform an omnidirectional scan of the surface of the mold cavity using a machine vision system to obtain the three-dimensional image data of the mold cavity;

[0159] Input the three-dimensional image data of the cavity into the image analysis detection model, and extract the features of the three-dimensional image data of the cavity using the deep convolutional neural network layer in the image analysis detection model to obtain the local features of the cavity;

[0160] Extract the features of the three-dimensional image data of the cavity using the graph neural network layer in the image analysis detection model to obtain the cavity topological structure features;

[0161] Connect the local features of the cavity and the cavity topological structure features using the fully connected layer in the image analysis detection model to obtain the first analysis and detection result output by the image analysis detection model;

[0162] Based on the first analysis and detection result, call a spectral analyzer to detect the material to be injection-molded to obtain the material property detection result of the material to be injection-molded;

[0163] Determine the material processing parameters of the material to be injection-molded according to the material property detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt;

[0164] Determine the first mold injection processing parameters based on the mold characteristic parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection processing parameters.

[0165] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

[0166] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0167] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or equivalently replace some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A mold processing and injection molding method based on machine vision, characterized in that, Including: Performing an omni-directional scan on the surface of the mold cavity based on a machine vision system to obtain the three-dimensional image data of the mold cavity; Inputting the three-dimensional image data of the mold cavity into the image analysis and detection model, and performing feature extraction on the three-dimensional image data of the mold cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain local features of the mold cavity; Performing feature extraction on the three-dimensional image data of the mold cavity based on the graph neural network layer in the image analysis and detection model to obtain topological structure features of the mold cavity; Connecting the local features of the mold cavity and the topological structure features of the mold cavity based on the fully connected layer in the image analysis and detection model to obtain the first analysis and detection result output by the image analysis and detection model; Based on the first analysis and detection result, calling a spectral analyzer to detect the material to be injection-molded to obtain the material property detection result of the material to be injection-molded; Determining the material processing parameters of the material to be injection-molded according to the material property detection result, and melting the material to be injection-molded based on the material processing parameters to obtain a material melt; Determining the first mold injection-molding processing parameters based on the mold property parameters of the mold cavity, and injecting the material melt into the mold cavity based on the first mold injection-molding processing parameters.

2. The injection molding method for mold processing based on machine vision according to claim 1, characterized in that, The mold property parameters include the heat capacity of the mold, the water channel diameter, the water channel length and the water channel spacing, as well as the material thermal conductivity, the filling volume and the minimum wall thickness of the mold cavity; determining the first mold injection-molding processing parameters based on the mold property parameters of the mold cavity includes: Obtaining the temperature change range of the mold cavity during the injection molding process and the heat exchange rate between the mold cavity and the surrounding environment; Calculating the heat balance time based on the temperature change range, the heat exchange rate and the heat capacity, and determining the injection molding cycle of the mold cavity based on the heat balance time; Determining the injection molding speed of the mold cavity based on the filling volume and the injection molding cycle; Determining the injection molding pressure of the mold cavity based on the minimum wall thickness and the injection molding speed; Determining the injection molding cycle, the injection molding speed and the injection molding pressure as the first mold injection-molding processing parameters.

3. The injection molding method for mold processing based on machine vision according to claim 1, wherein The step of calling a spectral analyzer to detect the material to be injection-molded based on the first analysis and detection result to obtain the material property detection result of the material to be injection-molded includes: If it is determined based on the first analysis and detection result that the surface of the mold cavity meets the preset result, then calling the spectral analyzer to emit multiple optical bands to detect the material to be injection-molded, and obtaining the reflectivity, absorptivity and transmittance of the material to be injection-molded for each optical band emitted by the spectral analyzer; Establishing a relationship model between the material properties of the material to be injection-molded and the optical bands of the spectral analyzer based on the reflectivity, absorptivity and transmittance of the material to be injection-molded for each optical band emitted by the spectral analyzer; Analyzing the relationship model to obtain the material property detection result of the material to be injection-molded; the material property detection result includes material composition information, material purity information and material physical property parameters.

4. The injection molding method for mold processing based on machine vision according to claim 1, wherein The machine vision system includes multiple laser emitters and a high-resolution CCD camera; Based on the machine vision system, a full-range scan of the surface of the mold cavity is performed to obtain the three-dimensional image data of the mold cavity, including: Obtaining the angle θ between the laser emitter and the surface of the mold cavity, the angle α between the high-resolution CCD camera and the laser beam emitted by the laser emitter, the displacement Δx of the laser spot formed on the surface of the mold cavity on the imaging plane of the high-resolution CCD camera, and the focal length f of the high-resolution CCD camera; Based on the angle θ, the angle α, the displacement Δx, and the focal length f, determining the height h of any point on the surface of the mold cavity, and the calculation formula for the height h is h = (Δx * f) / (tanθ * tanα); Processing and stitching the images of the surface of the mold cavity collected by the high-resolution CCD camera from multiple perspectives, calculating the surface height information of the mold cavity corresponding to each pixel point, and constructing the three-dimensional image data of the mold cavity.

5. The method for mold processing and injection molding based on machine vision according to any one of claims 1 to 4, characterized in that, The method further includes: During the injection of the material melt into the mold cavity with the injection processing parameters of the first mold, monitoring the flow state information of the material melt in the mold cavity; the flow state information includes the melt front position, the melt flow velocity, and the melt filling uniformity; Adjusting the injection processing parameters of the first mold based on the flow state information to obtain the injection processing parameters of the second mold; Injecting the material melt into the mold cavity based on the injection processing parameters of the second mold.

6. The mold processing and injection molding method based on machine vision according to any one of claims 1 to 4, characterized in that, The method further includes: After the mold injection is completed and before the product is demolded, based on the machine vision system, a comprehensive inspection of the injection-molded product of the mold is performed to obtain the three-dimensional image data of the injection-molded product of the mold; Inputting the three-dimensional image data of the product into the image analysis and detection model to obtain the second analysis and detection result output by the image analysis and detection model; Performing demolding treatment on the injection-molded product of the mold based on the second analysis and detection result.

7. An injection molding device for mold processing based on machine vision, characterized in that, Applied to the mold processing and injection method based on machine vision according to any one of claims 1 to 6; The mold processing and injection device based on machine vision includes: An image acquisition module for performing a full-range scan of the surface of the mold cavity based on the machine vision system to obtain the three-dimensional image data of the mold cavity; A model detection module for: Inputting the three-dimensional image data of the mold cavity into the image analysis and detection model, extracting features of the three-dimensional image data of the mold cavity based on the deep convolutional neural network layer in the image analysis and detection model to obtain local features of the mold cavity; Extracting features of the three-dimensional image data of the mold cavity based on the graph neural network layer in the image analysis and detection model to obtain topological structure features of the mold cavity; Connecting the local features of the mold cavity and the topological structure features of the mold cavity based on the fully connected layer in the image analysis and detection model to obtain the first analysis and detection result output by the image analysis and detection model; A material detection module, configured to call a spectral analyzer to detect the material to be injection-molded based on the first analysis and detection result, so as to obtain the material characteristic detection result of the material to be injection-molded; A processing module, configured to determine the material processing parameters of the material to be injection-molded according to the material characteristic detection result, and melt the material to be injection-molded based on the material processing parameters to obtain a material melt; An injection-molding module, configured to determine first mold injection-molding processing parameters based on the mold characteristic parameters of the mold cavity, and inject the material melt into the mold cavity based on the first mold injection-molding processing parameters.

8. An electronic device, comprising: A memory, configured to store a computer software program; A processor, configured to read and execute the computer software program, wherein when the processor executes the computer software program, the method for mold processing and injection-molding based on machine vision according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium storing a computer software program therein, characterized in that, When the computer software program is executed by the processor, the method for mold processing and injection-molding based on machine vision according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that, When the computer software program is executed by the processor, the method for mold processing and injection-molding based on machine vision according to any one of claims 1 to 6 is implemented.

Images