A method and system for real-time identification of dry-mix mortar production defects

Abstract

The application provides a dry-mixed mortar production defect real-time identification method and system. In the method, a multi-angle light field is projected to moving dry-mixed mortar by an annular light source array, the wavelength and angle of incident light are adjusted to form a high-contrast feature of aggregate edges; multi-spectral reflection data are collected, a characteristic waveband is separated based on the characteristic that the reflection intensity of aggregate is higher than that of cementitious material, and a reflection feature vector is generated; an aggregate distribution topology is constructed according to the vector space distribution; the topology is updated in structure feature transmission and connection strength, and when the connection strength is lower than a threshold value, the segregation boundary is determined and the closed area is segmented; the aggregate area proportion in the boundary and the inter-distance divergence generated quantity are calculated to generate a quantitative feature, and the stirring speed and cementitious ratio are cooperatively adjusted in combination with the material speed. The application realizes millisecond-level online identification and closed-loop suppression of dry-mixed mortar segregation risk, improves the uniformity of aggregate dispersion, and blocks the risk of hardening layer analysis.

Description

Technical Field

[0001] This application relates to the field of real-time defect identification technology, and in particular to a method and system for real-time identification of defects in dry mortar production. Background Technology

[0002] In dry-mix mortar production, fluctuations in raw material proportions (such as abnormal gradation of manufactured sand or excessive stone powder content) and deviations in mixing processes can easily lead to defects in the finished product, such as segregation, insufficient workability, and cracking, directly affecting construction quality and project safety. Traditional manual sampling and testing (such as sieving and viscosity testing) are insufficient to cover the real-time monitoring needs of high-speed continuous production lines (capacity ≥ 50 tons / hour). Especially when manufactured sand replaces natural sand, its complex and variable porosity, specific surface area, and stone powder viscosity significantly exacerbate the risk of quality fluctuations. Simultaneously, dust interference in the production environment and similar material colors (such as cement and stone powder) further increase the difficulty of defect identification. Therefore, there is an urgent need for a real-time detection technology capable of capturing the uniformity of material distribution, abnormal particle size, and mixing state online, and achieving dynamic closed-loop control of production parameters to ensure the workability and strength stability of the mortar.

[0003] Current targeted solutions include machine vision-based online particle size analysis systems. These systems capture real-time images of materials on the conveyor belt using high-resolution industrial cameras, combine morphological segmentation algorithms (such as the watershed algorithm) to identify aggregate agglomeration or segregation areas, and quantify particle distribution uniformity using grayscale statistics and texture analysis. Integrated at the mixer outlet or packaging section, these systems can be linked to a control console to adjust mixing time or additive feeding ratios. However, their core weakness lies in their low ability to distinguish impurities of similar color (such as fly ash adulterants) and their susceptibility to dust obscuring and uneven lighting, leading to a higher rate of missed detections of critical defects (such as localized stone powder enrichment). Summary of the Invention

[0004] This application provides a method and system for real-time identification of defects in dry mortar production, which solves the problems of lagging segregation boundary identification, mismatch in aggregate dispersion control, and failure of chromatographic blocking in the prior art.

[0005] Firstly, this application provides a method for real-time identification of defects in dry mortar production, including: A multi-angle incident light field is projected onto the surface of a continuously moving dry powder mortar material by a ring light source array, and the difference in spectral reflectance characteristics between aggregates and cementitious materials in the material is obtained. The incident wavelength and irradiation angle of the multi-angle incident light field are adjusted according to the difference in spectral reflectance characteristics to form high-contrast optical features at the edge of the aggregate. Multispectral reflectance data under the high-contrast optical characteristics are collected, and based on the characteristic that the reflectance intensity value of aggregate is higher than that of cementitious material, the characteristic bands corresponding to aggregate are separated from the multispectral reflectance data to generate the reflectance feature vectors of aggregate and cementitious material. The spatial distribution relationship of the reflection feature vector is obtained, and the distribution topology map of aggregate on the conveyor belt is constructed according to the spatial distribution relationship. The nodes in the distribution topology map represent aggregate regions, and the edges in the distribution topology map represent the spacing between adjacent aggregate regions. The topological structure features of the distributed topology map are transferred to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary, and the closed boundary of the separation region is segmented. The area ratio and spacing dispersion of the aggregate region within the closed boundary are calculated to generate the quantitative characteristics of the segregation region. At the same time, the material movement speed is obtained, and the coordinated adjustment of the stirring speed and the proportion of cementitious materials is triggered based on the mapping relationship between the quantitative characteristics and the material movement speed.

[0006] Optionally, the topological structure features of the distributed topology map are transferred to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary, and the closed boundary of the separation region is segmented, including: Extract the location coordinates of each node and the actual area of ​​the aggregate region corresponding to the node from the distribution topology map; Starting from each node, the actual area of ​​the aggregate region is transmitted to the adjacent nodes, and the difference between the actual area of ​​the aggregate region of the node and the adjacent nodes is calculated. The difference is then weighted and fused with the edge length value to obtain the connection strength value of the edge. Traverse the connection strength values ​​of all edges in the distribution topology graph. When the connection strength value is lower than a preset threshold, mark the edge corresponding to the connection strength value as a segregation boundary that characterizes the abnormal distribution of aggregate. Connect all edges marked as separation boundaries to form a closed boundary contour with a continuous path, and the region of the nodes enclosed by the closed boundary contour constitutes the closed boundary of the separation region.

[0007] Optionally, a multi-angle incident light field is projected onto the surface of a continuously moving dry mortar material using a ring light source array, and the difference in spectral reflectance characteristics between the aggregate and cementitious material in the material is obtained. The incident wavelength and irradiation angle of the multi-angle incident light field are adjusted according to the difference in spectral reflectance characteristics to form high-contrast optical features at the aggregate edge, including: By simultaneously projecting incident light fields at different irradiation angles onto the surface of continuously moving dry powder mortar material through a ring light source array, the reflected light data of the material surface is obtained at the same time. Extract the aggregate reflection intensity value and the cementitious material reflection intensity value from the reflected light data, and extract the difference in spectral reflection characteristics between the aggregate and the cementitious material from the reflected light data. The difference in spectral reflection characteristics is that the reflection intensity value of the aggregate is higher than that of the cementitious material. Based on the differences in spectral reflectance characteristics, the incident light wavelength of the ring light source array and the illumination angle of each angle light source are adjusted in real time. The incident light wavelength and irradiation angle are adjusted and re-projected onto the material surface, so that the intensity of reflected light in the aggregate edge area is significantly higher than that in the cementitious material area, thus forming a high-contrast optical feature at the aggregate edge.

[0008] Optionally, the area ratio and spacing dispersion of the aggregate region within the closed boundary are calculated to generate the quantitative characteristics of the segregation region. Simultaneously, the material movement velocity is obtained, and based on the mapping relationship between the quantitative characteristics and the material movement velocity, the coordinated adjustment of the stirring speed and the proportion of cementitious materials is triggered, including: Within the closed boundary of the segregation region, the total area of ​​all aggregate regions is calculated, and the area ratio of the aggregate regions is obtained based on the total area and the total area of ​​the closed boundary. Extract the average value of the edge length between all adjacent aggregate region nodes within the closed boundary, and calculate the deviation coefficient of the edge length value from the average value as the spacing dispersion; A shaft speed sensor is installed on the rear of the conveyor belt to obtain the current material movement speed in real time. At the same time, the area ratio value and the spacing dispersion are combined into a quantitative feature structure. Based on the preset mapping relationship between the material movement speed and the quantified feature structure, the target value offset of the current stirring speed and the component increment value of the cementitious material ratio are determined. The target value offset is converted into a frequency control command for the stirring motor, and the component increment value is converted into an opening control command for the feeding valve to trigger the coordinated adjustment action of the execution equipment.

[0009] Optionally, the spatial distribution relationship of the reflection feature vectors is obtained, and a distribution topology map of the aggregate on the conveyor belt is constructed based on the spatial distribution relationship. Nodes in the distribution topology map represent aggregate regions, and edges in the distribution topology map represent the spacing between adjacent aggregate regions, including: Obtain the reflection feature vectors of all observation units on the conveyor belt, and extract the average component of the reflection intensity value of the aggregate in the reflection feature vectors. The average component represents the content level of aggregate in each observation unit. The position coordinates of the observation unit on the conveyor belt are located, and a mapping relationship between the position coordinates and the aggregate content level is established. The mapping relationship constitutes the spatial distribution relationship of the reflection feature vector. Extract continuous observation unit groups where the aggregate content level on the conveyor belt exceeds a preset threshold, and use the center position of each continuous observation unit group as a node in the distribution topology map, where the node represents the actual distribution location of the aggregate area; Calculate the straight-line distance between any two adjacent aggregate regions and use the straight-line distance as the edge length value connecting the two nodes in the distribution topology graph; The actual distribution locations and side length values ​​are combined into a mesh structure to serve as a distribution topology map characterizing the distribution of aggregates on the conveyor belt.

[0010] Optionally, the position coordinates of the observation unit on the conveyor belt are located, and a mapping relationship between the position coordinates and the aggregate content level is established. This mapping relationship constitutes the spatial distribution relationship of the reflection feature vector, including: Positioning markers are set at equal intervals on both sides of the conveyor belt. When the observation unit moves with the conveyor belt and passes the positioning markers, it triggers the synchronization pulse signal of the optical sensor. The position coordinates of each observation unit in the conveyor belt are determined by matching the timestamp of the synchronization pulse signal with the count value of the rotary encoder of the conveyor belt. A mapping relationship is established between the actual location coordinates of each observation unit and the content level of the aggregate to form a data unit that combines location and content. Collect data units of location and content combinations of all observation units, and construct a spatial distribution relationship describing the aggregate distribution state on the material surface according to the location coordinates.

[0011] Optionally, multispectral reflectance data under the high-contrast optical characteristics are collected, and based on the characteristic that the reflectance intensity value of the aggregate is higher than that of the cementitious material, the characteristic bands corresponding to the aggregate are separated from the multispectral reflectance data to generate reflectance feature vectors of the aggregate and the cementitious material, including: The reflected light from the surface of a material under high-contrast optical characteristics is collected by an optical sensor to obtain multispectral reflectance data containing multiple spectral band reflectance intensity values; Based on the characteristic that the reflection intensity value of aggregate is always higher than that of cementitious material, a continuous spectral band range that meets the above characteristic is selected from the complete spectral bands of multispectral reflectance data, and the continuous spectral band range is determined as the characteristic band corresponding to the aggregate. The characteristic band of the aggregate and the range outside the characteristic band are combined to form a reflection structure, and the reflection structure is defined as the reflection characteristic vector of the aggregate and cementitious material.

[0012] Secondly, this application provides a real-time defect identification system for dry mortar production, including: The adjustment module is used to project a multi-angle incident light field onto the surface of a continuously moving dry powder mortar material through a ring light source array, and to obtain the difference in spectral reflectance characteristics between aggregates and cementitious materials in the material. Based on the difference in spectral reflectance characteristics, the incident wavelength and irradiation angle of the multi-angle incident light field are adjusted to form a high-contrast optical feature at the edge of the aggregate. The separation module is used to collect multispectral reflectance data under the high-contrast optical features, and based on the characteristic that the reflectance intensity value of the aggregate is higher than that of the cementitious material, it separates the characteristic bands corresponding to the aggregate from the multispectral reflectance data to generate the reflectance feature vectors of the aggregate and the cementitious material. The acquisition module is used to acquire the spatial distribution relationship of the reflection feature vector and construct a distribution topology map of aggregates on the conveyor belt based on the spatial distribution relationship. The nodes in the distribution topology map represent aggregate regions, and the edges in the distribution topology map represent the spacing between adjacent aggregate regions. The update module is used to transfer topological structure features to the distribution topology map in order to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary and the closed boundary of the separation region is segmented. The triggering module is used to calculate the area ratio and spacing dispersion of the aggregate region within the closed boundary to generate the quantitative characteristics of the segregation region. At the same time, it acquires the material movement speed and triggers the coordinated adjustment of the stirring speed and the proportion of cementitious materials based on the mapping relationship between the quantitative characteristics and the material movement speed.

[0013] Thirdly, embodiments of this application provide a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a real-time identification method for dry mortar production defects as described in the first aspect above.

[0014] Fourthly, embodiments of this application provide a computer storage medium storing a computer program, which, when executed by a computer, implements a real-time identification method for dry mortar production defects as described in the first aspect.

[0015] In this application's technical solution, intelligent detection and control of dry mortar segregation are achieved through multispectral optical feature analysis and topological structure calculation. Specifically, high-contrast optical features based on multi-angle incident light field adjustment significantly improve the accuracy of aggregate edge recognition; spatial distribution analysis of reflection feature vectors accurately constructs the topological relationship of aggregate distribution; and a topological structure feature transfer algorithm enables automatic segmentation and quantitative evaluation of segregation boundaries. This method overcomes the lag defects of traditional manual sampling inspection, enabling real-time monitoring of aggregate segregation during transportation. Dynamic and coordinated adjustment of stirring speed and material ratio effectively ensures the uniformity of dry mortar, significantly improving the production quality and construction performance of building materials.

[0016] Furthermore, high-precision identification of the optical features of dry mortar material surfaces is achieved through dynamic adjustment of the ring light source array. Spectral characteristic identification based on multi-angle synchronous projection and real-time reflection analysis significantly improves the distinguishability between aggregates and cementitious materials; the adaptive adjustment mechanism of incident light wavelength and illumination angle effectively enhances the optical contrast of aggregate edges; and the generation of high-contrast optical features provides high-quality input data for subsequent segregation detection. This method overcomes the limitations of traditional fixed light source detection, achieving real-time dynamic identification of material components, providing a reliable optical detection basis for dry mortar uniformity monitoring, and significantly improving the intelligence level of building material production processes.

[0017] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart of a real-time identification method for dry mortar production defects provided in this application is shown; Figure 2 A scene diagram is shown illustrating a real-time identification method for dry mortar production defects provided in this application; Figure 3 This invention provides a schematic diagram of the structure of a real-time defect identification system for dry mortar production. Figure 4 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0021] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0022] Researchers have discovered a fundamental bottleneck in the quality monitoring of dry-mix mortar production: while machine vision-based particle size analysis can identify aggregate agglomeration, its failure in color discrimination and lack of environmental resistance lead to a double loss of control. Specifically, spectral confusion between cement and similar-colored materials like stone powder results in a high rate of missed detection for stone powder enrichment, and dust obscuring and uneven lighting cause misjudgments of segregation boundaries. This contradiction stems from the limitations of traditional technologies in decoupling from differences in material optical properties and their passive anti-interference mechanisms, necessitating the construction of a real-time control architecture that combines multispectral active sensing and topological quantification analysis.

[0023] To address the aforementioned challenges, this invention proposes a real-time defect identification method for dry mortar production. Its innovation lies in breaking through the limitations of traditional visual inspection through high-contrast optical feature capture and a dynamic topological decision chain. Specifically, a ring light source array projects a multi-angle adjustable light field. Based on the difference in reflectivity between aggregates and cementitious materials, the wavelength and incident angle are dynamically adjusted to generate enhanced spectral features at aggregate edges. Aggregate characteristic bands are separated from multispectral data to construct a spatial distribution topology map (nodes represent aggregate regions, edges represent region spacing). The node connection strength is updated through topological structure feature transmission. Separation boundaries are segmented using connection strength thresholds, and the area ratio and spacing dispersion within closed regions are calculated. Finally, by combining the material movement speed mapping relationship, the coordinated adjustment of stirring speed and cementitious material ratio is triggered in real time. This method breaks through traditional limitations: the multispectral active light field achieves nanoscale separation of the reflective features of manufactured sand and stone powder for the first time, completely solving the problem of optical confusion between materials with similar colors; the topological boundary segmentation algorithm accurately locates the segregation region at a preset flow rate by dynamically updating the connection strength; the closed-loop control chain adjusts the proportioning parameters in real time according to the discrete quantization characteristics, forming a collaborative decision flow of "spectral decoupling - topological quantization - boundary segmentation - proportioning control", which significantly improves the anti-segregation performance and workability index.

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Figure 1 This application provides a flowchart of a method for real-time identification of defects in dry mortar production, as illustrated in the embodiments of this application. Figure 1 As shown, the method includes: 101. A multi-angle incident light field is projected onto the surface of a continuously moving dry powder mortar material through a ring light source array, and the difference in spectral reflectance characteristics between the aggregate and cementitious material in the material is obtained. The incident wavelength and irradiation angle of the multi-angle incident light field are adjusted according to the difference in spectral reflectance characteristics to form a high-contrast optical feature at the edge of the aggregate.

[0026] Optionally, step 101 may specifically include the following steps: 1011. By using a ring light source array to synchronously project incident light fields at different irradiation angles onto the surface of continuously moving dry powder mortar material, and at the same time acquire the reflected light data of the material surface.

[0027] 1012. Extract the aggregate reflection intensity value and the cementitious material reflection intensity value from the reflected light data, and extract the difference in spectral reflection characteristics between the aggregate and the cementitious material from the reflected light data. The difference in spectral reflection characteristics is manifested in that the reflection intensity value of the aggregate is higher than that of the cementitious material.

[0028] 1013. Based on the differences in spectral reflectance characteristics, adjust the incident light wavelength of the ring light source array and the illumination angle of each angle light source in real time.

[0029] 1014. The adjusted incident light wavelength and irradiation angle are re-projected onto the material surface, so that the intensity of reflected light in the aggregate edge area is significantly higher than that in the cementitious material area, thereby forming a high-contrast optical feature at the aggregate edge.

[0030] In the above scheme, a ring light source array refers to a combination of multi-angle light sources arranged in a ring. Dry powder mortar refers to a powdery building material composed of aggregates and cementitious materials. A multi-angle incident light field refers to a combination of light fields irradiated from different angles. Differences in spectral reflectance characteristics refer to the differences in light reflectance characteristics of different materials. Aggregate edge refers to the boundary region of aggregate particles. High-contrast optical characteristics refer to optical signals with significant intensity differences. Reflected light data refers to data recording the light reflection characteristics of the material surface. Aggregate reflectance intensity value refers to the intensity value of light reflected by the aggregate. Cementitious material reflectance intensity value refers to the intensity value of light reflected by the cementitious material. Incident light wavelength refers to the wavelength parameter of the projected light. Irradiation angle refers to the angle between the light ray and the material surface.

[0031] In this embodiment, firstly, through step 1011, a ring-shaped light source array (composed of multiple independently adjustable LED units with adjustable angles and wavelengths) projects a multi-directional incident light field onto the continuously moving surface of dry powder mortar material. The array synchronously adjusts the illumination angle of each light source unit via a motor-driven mechanism, ensuring that light covers the material surface from different directions. Simultaneously, a high-speed industrial camera captures reflected light data in real time while the material is in motion. This data includes the distribution of reflection intensity of aggregates and cementitious materials under illumination at different angles. This process draws on multi-angle uniform illumination technology from machine vision, using an aspherical plano-convex lens to constrain the light divergence angle, avoiding stray light interference and providing raw optical signals for subsequent analysis.

[0032] Subsequently, in step 1012, the system performs spectral feature separation processing on the collected reflected light data: Reflection intensity value extraction: The reflection signals of aggregates (mainly composed of high-reflectivity minerals such as quartz and feldspar) and cementitious materials (low-reflectivity materials such as cement) are separated using a spectral dispersion algorithm. Due to their dense crystal structure and the presence of metal oxides, aggregates generally have higher reflection intensity values ​​than cementitious materials; Quantification of spectral reflection characteristic differences: The difference in reflection intensity between the two types of materials is calculated using a differential enhancement algorithm to generate a difference feature matrix. This difference is manifested in the fact that the reflectivity of aggregates is significantly higher than that of cementitious materials in a specific wavelength band, forming a quantifiable spectral boundary marker.

[0033] Next, in step 1013, the system adjusts the light source parameters in real time through a dynamic feedback control module based on the spectral reflectance difference matrix: incident light wavelength adjustment: strengthening the sensitive wavelength band of aggregates and suppressing the response wavelength band of cementitious materials; irradiation angle optimization: based on the high surface roughness of aggregates (leading to directional reflection), the system controls the light source units at specific angles in the annular light source array to increase the incident angle (e.g., from 30° to 60°), utilizing the principle of specular reflection enhancement to improve the reflection intensity at the aggregate edges. This process dynamically matches the optimal combination of light source parameters through a PID control algorithm.

[0034] Finally, in step 1014, the system reprojects the optimized incident light wavelength and illumination angle onto the material surface. The adjusted light field causes specular reflection in the aggregate edge region (due to geometric angular effects), increasing the intensity of reflected light compared to the flat cementitious material region; while the cementitious material, due to its rough surface and high porosity, mainly forms diffuse reflection, significantly reducing the reflection intensity. The contrast difference between the two is widened, thus forming a clear, high-contrast optical feature. This feature, captured by an industrial camera, can be directly used for automatic identification of aggregate particle size distribution and analysis of mixing uniformity.

[0035] In practical applications, in the conveyor belt stage of a dry mortar production line, the system uses a ring light source array to simultaneously project incident light fields at different irradiation angles (such as 15°, 45°, and 75° incident light) onto the surface of the continuously moving dry mortar material (the material passes through the detection area at a uniform speed). At the same time, a high-speed industrial camera acquires the reflected light data of the material surface (covering the mixed area of ​​aggregate and cementitious material).

[0036] The system extracts the reflection intensity values ​​of aggregates and cementitious materials from the reflected light data (aggregates appear as bright spots due to the high reflectivity of quartz components, while cementitious materials appear as dark areas due to the low reflectivity of cement components), and extracts the difference in spectral reflection characteristics between aggregates and cementitious materials from the reflected light data based on a spectral analysis algorithm (this difference is manifested in the fact that the reflection intensity value of aggregates is always higher than that of cementitious materials).

[0037] To optimize the accuracy of aggregate edge recognition, the system adjusts the incident light wavelength of the ring light source array in real time (switching the wavelength to the infrared band that aggregates are sensitive to) and the illumination angle of each light source (increasing the angle of the side light source to enhance the shadow contrast of the aggregate outline) based on the differences in spectral reflectance characteristics (such as when the difference in reflectance intensity between aggregate and cementitious material is lower than the threshold).

[0038] Subsequently, the system reprojects the adjusted incident light wavelength and illumination angle onto the material surface (e.g., infrared band + 75° side light). This causes the reflected light intensity at the aggregate edge area to be significantly higher than that at the cementitious material area due to the prism effect and shadow superposition (the cementitious material maintains low reflectivity due to its light absorption properties), creating a high-contrast optical feature at the aggregate edge. This optical feature is captured by an industrial camera and transmitted to a defect recognition algorithm to accurately locate segregation defects caused by aggregate agglomeration or cementitious material deficiency (e.g., blurred areas at aggregate edges indicate uneven mixing). This drives the production line to automatically adjust the mixer speed or raw material ratio to meet the process standards for the uniformity of dry mortar.

[0039] The overall solution in step 101 above achieves intelligent enhancement and precise control of the optical features of dry mortar material surfaces. By projecting incident light fields from multiple angles using a ring light source array, combined with the analysis of the differences in spectral reflectance characteristics between aggregates and cementitious materials, the incident light wavelength and illumination angle are dynamically adjusted, innovatively forming high-contrast optical features at the aggregate edges. This technology overcomes the limitations of traditional uniform illumination, providing a high signal-to-noise ratio visual inspection basis for subsequent material component identification, and significantly improving the recognizability of material surface features.

[0040] 102. Collect multispectral reflectance data under the high-contrast optical characteristics, and based on the characteristic that the reflectance intensity value of the aggregate is higher than that of the cementitious material, separate the characteristic bands corresponding to the aggregate from the multispectral reflectance data to generate the reflectance feature vectors of the aggregate and the cementitious material.

[0041] Optionally, step 102 may specifically include the following steps: 1021. The reflected light from the surface of a material under high-contrast optical characteristics is collected by an optical sensor to obtain multispectral reflectance data containing multiple spectral band reflectance intensity values.

[0042] 1022. Based on the characteristic that the reflection intensity value of aggregate is always higher than that of cementitious material, a continuous spectral band range that meets the aforementioned characteristic is selected from the complete spectral bands of multispectral reflectance data, and the continuous spectral band range is determined as the characteristic band corresponding to the aggregate.

[0043] 1023. The characteristic band of the aggregate and the range outside the characteristic band are combined to form a reflection structure, and the reflection structure is defined as the reflection characteristic vector of the aggregate and cementitious material.

[0044] In the above scheme, multispectral reflectance data refers to a data set containing reflectance intensities across multiple wavelength bands. An optical sensor is a sensing device that detects optical signals. Reflected light refers to light reflected from the surface of a material. A spectral band reflectance intensity value refers to the intensity of reflected light within a specific wavelength range. A complete spectral band refers to the entire spectral range detected by the sensor. A continuous spectral band range refers to a continuous wavelength interval within the spectrum. Reflection structure refers to the organizational form of reflection characteristics. A reflection eigenvector is a mathematical vector characterizing the reflectance properties of a material. A characteristic band refers to a band where the reflection characteristics of a specific material are prominent.

[0045] In this embodiment, firstly, in step 1021, the system uses an optical sensor (such as a multispectral imaging device) to align with the material surface under high-contrast optical features. These features are generated by the light field optimization in step 101, making the reflection intensity at the aggregate edge significantly higher than that of the cementitious material. The optical sensor decomposes the incident light into reflected light of different wavelengths using a beam splitter or filter array, and the CCD or CMOS photosensitive element simultaneously captures the light intensity signals of each spectral band. The sensor acquires the spectral response of the dynamic material surface using line scanning or area array imaging, generating a multispectral reflectance data matrix containing tens to hundreds of spectral band reflection intensity values. The rows and columns of this matrix correspond to the material's spatial coordinates and spectral channels, forming a three-dimensional data cube, providing the raw input for subsequent feature separation.

[0046] Subsequently, in step 1022, based on the inherent characteristic that the reflection intensity value of the aggregate is always higher than that of the cementitious material (due to the difference in near-infrared reflectivity caused by the density of mineral crystals), the system performs a band filtering algorithm on the multispectral reflectance data: it iterates through all bands of reflection intensity values, calculates the average reflection intensity difference between the aggregate region and the cementitious material region at each wavelength, and filters out continuous band ranges where the difference is consistently positive and exceeds the noise threshold. This process employs a sliding window verification mechanism to ensure that adjacent bands satisfy the reflection intensity order relationship, and finally determines the band range that meets the conditions as the characteristic band corresponding to the aggregate. This step achieves automatic spectral segmentation through differences in physical properties, eliminating interference from the cementitious material.

[0047] Finally, in step 1023, the system combines the characteristic bands of the aggregate (high-difference bands) with the range outside the characteristic bands (low-difference bands) into a reflection structure: the high-difference bands reflect the spatial distribution characteristics of the aggregates, and the low-difference bands characterize the background response of the cementitious materials. The two types of band data are normalized and integrated into a unified vector using a feature stitching algorithm, with each spatial point corresponding to a joint descriptor containing high / low-difference channels. This reflection structure is defined as the reflection feature vector of the aggregates and cementitious materials, with a dimension equal to the total number of spectral bands. The high-difference bands have a significantly higher weight than the low-difference bands in the vector, thus mathematically solidifying the separability of the two types of materials and providing input for subsequent mixing uniformity analysis or particle size statistics.

[0048] In practical applications, during the continuous production of crack-resistant mortar, the system illuminates the surface of the material passing at a uniform speed through a preset high-contrast optical feature (the light field conditions generated by the optimized ring light source array). At this time, the optical sensor (such as a multispectral industrial camera) simultaneously collects the reflected light from the surface of the material under the high-contrast optical feature (capturing the difference between the reflection peaks and valleys at the interface between aggregate and cementitious material), and obtains multispectral reflectance data (covering the visible to near-infrared bands) containing multiple spectral bands of reflectance intensity values.

[0049] Based on the inherent characteristic that the reflectance of aggregates is always higher than that of cementitious materials (the reflectance of quartz components in aggregates is significantly higher than that of cement-based cementitious materials in a specific wavelength band), the system filters out continuous spectral ranges that meet the aforementioned characteristics from the complete spectral bands of multispectral reflectance data (e.g., the spectral range in the near-infrared band where the reflectance of aggregates is consistently higher than that of cementitious materials), and determines the continuous spectral range as the characteristic band corresponding to the aggregates (where the difference in reflectance is most significant).

[0050] Subsequently, the system combines the characteristic wavelength band of the aggregate (near-infrared high reflectivity range) with the range outside the characteristic wavelength band (such as the visible light band) to form a reflection structure (this structure includes the dominant reflection peak of the aggregate characteristic wavelength band and the low reflectivity plateau outside the cementitious material characteristic wavelength band), and defines the reflection structure as the reflection characteristic vector of the aggregate and cementitious material. This vector is input into the defect identification model to determine the uniformity of aggregate distribution in real time (e.g., uneven distribution of reflection intensity in the characteristic wavelength band indicates aggregate segregation), driving the production line to adjust the mixing parameters to ensure that the crack-resistant mortar meets the homogeneity requirements for aggregate distribution.

[0051] The overall scheme in step 102 above achieves intelligent extraction and vectorized representation of aggregate reflectance characteristics. Based on multispectral reflectance data under high-contrast optical characteristics, and utilizing the difference in reflectance intensity between aggregates and cementitious materials, an innovative feature band screening algorithm is designed to construct a reflectance feature vector reflecting the material composition distribution. This technology realizes the intelligent conversion from raw spectral data to structured feature vectors, providing a quantifiable mathematical expression model for material distribution analysis.

[0052] 103. Obtain the spatial distribution relationship of the reflection feature vector, and construct the distribution topology map of aggregate on the conveyor belt based on the spatial distribution relationship. The nodes in the distribution topology map represent aggregate regions, and the edges in the distribution topology map represent the spacing between adjacent aggregate regions.

[0053] Optionally, step 103 may specifically include the following steps: 1031. Obtain the reflection feature vectors of all observation units on the conveyor belt, and extract the average value component of the reflection intensity value of the aggregate in the reflection feature vectors, wherein the average value component represents the content level of aggregate in each observation unit.

[0054] 1032. Locate the position coordinates of the observation unit on the conveyor belt, and establish a mapping relationship between the position coordinates and the aggregate content level. The mapping relationship constitutes the spatial distribution relationship of the reflection feature vector.

[0055] Step 1032 may specifically include the following process: Positioning markers are set at equal intervals on both sides of the conveyor belt. When an observation unit moves past the positioning markers with the conveyor belt, a synchronization pulse signal of the optical sensor is triggered. The timestamp of the synchronization pulse signal is matched with the count value of the rotary encoder of the conveyor belt to determine the position coordinates of each observation unit in the conveyor belt. A mapping relationship is established between the actual position coordinates of each observation unit and the aggregate content level to form a data unit combining position and content. All data units combining position and content of the observation units are collected, and a spatial distribution relationship describing the aggregate distribution state on the material surface is constructed according to the position coordinates.

[0056] 1033. Extract continuous observation unit groups where the aggregate content level on the conveyor belt exceeds a preset threshold, and use the center position of each continuous observation unit group as a node in the distribution topology map, where the node represents the actual distribution position of the aggregate area.

[0057] 1034. Calculate the straight-line distance between any two nodes in adjacent aggregate regions, and use the straight-line distance as the edge length value connecting the two nodes in the distribution topology graph.

[0058] 1035. The actual distribution locations and side length values ​​are combined into a mesh structure to form a distribution topology map that characterizes the distribution of aggregates on the conveyor belt.

[0059] In the above scheme, the spatial distribution relationship of the reflection feature vector refers to the positional distribution characteristics of the reflection features on the conveyor belt. The distribution topology map is a graphical representation reflecting the aggregate distribution relationship. The conveyor belt is a device used for continuous material transport. The observation unit is the smallest area for optical detection. The average component is the average value of the aggregate reflection intensity in the reflection feature vector. The content level refers to the proportion of aggregate in the observation unit. The position coordinates are the spatial position of the observation unit on the conveyor belt. The mapping relationship is the correspondence between position and content. The positioning marker is a physical marker used for position calibration. The synchronization pulse signal is a trigger signal synchronized with the position marker. The timestamp is a time stamp recording the moment the signal occurs. The rotary encoder is a device for measuring the movement position of the conveyor belt. The matching operation is the processing procedure that maps signals to positions. The data unit is a data combination containing position and content. The continuous observation unit group is a set of spatially adjacent observation units. The preset threshold is the critical value for judging aggregate content. The center position is the geometric center point of the continuous observation unit group. The node is a key position point in the distribution topology map. The edge length value is the distance between nodes. A mesh structure is a network graph consisting of nodes and edges.

[0060] In this embodiment, firstly, through step 1031, the system obtains the reflection feature vector of all observation units on the conveyor belt (generated in step 102). This vector contains the reflection intensity data of aggregate in multiple spectral bands. The system focuses on the average component in the vector through a vector component extraction algorithm, that is, it calculates the arithmetic mean of the aggregate reflection intensity values ​​of all spectral bands within each observation unit. This average component directly quantifies the overall reflection intensity level of the aggregate within the unit. Since the aggregate reflection intensity value is positively correlated with the content, this component is defined as the core indicator characterizing the content level of aggregate in each observation unit, providing basic data input for subsequent distribution analysis.

[0061] Subsequently, in step 1032, the system establishes a spatial mapping relationship using position synchronization calibration technology. The specific process is as follows: Positioning markers are installed at equal intervals on both sides of the conveyor belt. When an observation unit carrying material passes the markers with the conveyor belt, a synchronization pulse signal from the optical sensor is triggered. The system matches the timestamp of the synchronization pulse signal with the real-time count value of the conveyor belt's rotary encoder (each encoder pulse corresponds to a fixed displacement), and calculates the precise position coordinates of each observation unit from the start point of the conveyor belt using a displacement-time conversion formula. Next, the system binds the actual position coordinates of each observation unit to its corresponding aggregate content level, forming a data unit containing both position and content attributes. Finally, all data units are aggregated and sorted by position coordinates to construct a complete spatial distribution relationship of reflection feature vectors. This relationship is stored in matrix form, with row vectors representing the observation unit positions and column vectors storing content level values.

[0062] Next, in step 1033, the system performs aggregate enrichment region identification based on the data unit sequence in the spatial distribution relationship. The system scans the aggregate content level of all observation units, filters out segments where the content of multiple consecutive units exceeds a preset threshold, and marks such segments as continuous observation unit groups. For each continuous observation unit group, the system calculates the average position coordinates of all units within the group, uses this as the geometric center point of the aggregate enrichment region, and defines this center point as a node in the distribution topology map. The coordinate values ​​of this node represent the actual distribution location of the aggregate region, and its quantity, together with the spatial coordinates, constitutes the skeleton framework of the topology map.

[0063] Then, in step 1034, the system performs adjacency analysis on all nodes generated in step 1033. After sorting the nodes by their position coordinates along the conveyor belt direction, the system calculates the straight-line distance (two-dimensional Euclidean distance) between two adjacent nodes. This distance value is directly assigned as the edge length connecting the two nodes in the topology graph. This process is automatically completed by the geometric relationship analysis engine, ensuring that each edge accurately reflects the actual spacing of the aggregate area.

[0064] Finally, in step 1035, the system integrates the outputs of the previous steps: the actual distribution locations of the aggregate regions represented by the nodes and their edge lengths are input into the topology synthesis module. This module constructs a mesh structure with nodes as vertices and edge lengths as connection weights. This structure is based on graph theory, and the connections between nodes fully preserve the actual spatial adjacency. The edge length data reflects the real physical spacing, ultimately generating a complete distribution topology map that characterizes the distribution of aggregates on the conveyor belt. This map can be directly used for aggregate distribution uniformity analysis or gradation optimization decisions.

[0065] In practical applications, in the aggregate segregation monitoring system of the crack-resistant mortar production line, the system uses optical sensors to collect the reflection feature vector of the crack-resistant mortar material on the conveyor belt in real time (generated by high-contrast light field optimization), obtains the reflection feature vector of all observation units on the conveyor belt (the observation unit is an equally divided grid area on the surface of the conveyor belt), and extracts the average value component of the reflection intensity value of the aggregate in the reflection feature vector (this component directly represents the content level of quartz aggregate in each observation unit).

[0066] Next, to establish spatial mapping, equally spaced positioning markers (the spacing between the markers matches the width of the observation unit) are set on both sides of the conveyor belt. When the observation unit moves past the positioning markers with the conveyor belt, a synchronization pulse signal is triggered from the optical sensor. The system matches the timestamp of the synchronization pulse signal with the count value of the rotary encoder of the conveyor belt (the encoder records the displacement of the conveyor belt) to determine the position coordinates of each observation unit in the conveyor belt (accurate to millimeter-level grid coordinates). A mapping relationship is established between the actual position coordinates of each observation unit and the aggregate content level to form a data unit combining position and content. All data units combining position and content of the observation units are collected, and a spatial distribution relationship describing the aggregate distribution state on the material surface is constructed according to the position coordinates.

[0067] Based on this relationship, the system extracts continuous observation unit groups (such as high-content strip areas formed by aggregate aggregation) where the aggregate content level on the conveyor belt exceeds a preset threshold. The center position of each continuous observation unit group is used as a node in the distribution topology map (the node accurately marks the actual distribution location of the aggregate area). Subsequently, the straight-line distance between nodes of any two adjacent aggregate areas (such as the Euclidean distance between node A and node B) is calculated, and the straight-line distance is used as the edge length value connecting the two nodes in the distribution topology map (step 1034).

[0068] Ultimately, the system combines the actual distribution locations with the edge length values ​​to form a network structure (nodes are aggregate aggregation centers, and edge lengths are the distances between adjacent aggregation areas), serving as a distribution topology map characterizing the aggregate distribution on the conveyor belt. This topology map is input into the segregation defect identification model in real time: if the edge length value between adjacent nodes is too small (indicating dense aggregate) or too large (indicating enriched cementitious materials), it is determined to be an aggregate segregation defect, automatically triggering the mixer speed adjustment or raw material compensation mechanism to ensure that the crack-resistant mortar meets the homogeneity requirements of aggregate distribution in GB / T25181.

[0069] The overall scheme in step 103 above achieves spatial topological modeling and visualization of aggregate distribution. Through spatial distribution relationship analysis of reflection feature vectors, it innovatively transforms the aggregate distribution on the conveyor belt into a node-edge structure topological graph, where nodes represent aggregate aggregation areas and edges represent aggregate spacing relationships. This technology realizes the abstract transformation of material distribution from physical space to a mathematical network, providing an intuitive and computable spatial relationship model for segregation detection.

[0070] 104. Transfer topological structure features to the distributed topology map to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary, and the closed boundary of the separation region is segmented.

[0071] Optionally, step 104 may specifically include the following steps: 1041. Extract the position coordinates of each node and the actual area of ​​the aggregate region corresponding to the node from the distribution topology map.

[0072] 1042. Starting from each node, the actual area of ​​the aggregate region is transmitted to the adjacent nodes, and the difference between the actual area of ​​the aggregate region of the node and the adjacent nodes is calculated. The difference is then weighted and fused with the edge length value to obtain the connection strength value of the edge.

[0073] 1043. Traverse the connection strength values ​​of all edges in the distribution topology graph. When the connection strength value is lower than a preset threshold, mark the edge corresponding to the connection strength value as a segregation boundary that represents the abnormal distribution of aggregate.

[0074] 1044. Connect all edges marked as separation boundaries to form a closed boundary contour with a continuous path, and form the closed boundary of the region of nodes enclosed by the closed boundary contour as the closed boundary of the separation region.

[0075] In the above scheme, topology feature transfer refers to the calculation process of transferring node features along the edges. Connection strength is a quantitative indicator of the degree of association between nodes. A closed boundary refers to a completely closed boundary profile. Location coordinates refer to the spatial position of a node on the conveyor belt. Actual area refers to the actual coverage area of ​​the aggregate region. Difference refers to the difference in area between two aggregate regions. Weighted fusion calculation refers to a calculation method that weights multiple parameters. Connection strength value refers to the numerical value that quantifies the connection strength of nodes. Preset threshold refers to the critical value for judging connection strength. Abnormal aggregate distribution refers to the phenomenon of uneven aggregate distribution. A closed boundary profile refers to a completely closed boundary line. Segregation region refers to a region where aggregate distribution is discontinuous.

[0076] In this embodiment, firstly, through step 1041, the system extracts the position coordinates (i.e., the two-dimensional spatial position of the aggregate region on the conveyor belt) of each node from the distribution topology map (generated in step 103) and the actual area of ​​the aggregate region corresponding to that node (calculated by image segmentation algorithm to calculate pixel area and convert it to physical area). Specifically, the process involves traversing all nodes in the topology map, calling the position coordinate data (such as x and y values ​​in a Cartesian coordinate system) stored in the node attributes, and simultaneously obtaining the actual coverage area of ​​the aggregate region represented by that node on the material surface (based on pixel statistics and spatial calibration conversion from hyperspectral imaging). These two types of data constitute the basic physical attribute set of the nodes, providing spatial and geometric feature inputs for subsequent connection strength calculations.

[0077] Subsequently, in step 1042, the system performs topology feature transfer: starting from each node, the actual area data of the aggregate region is transferred to adjacent nodes along the edges of the topology graph. The transfer process uses a neighborhood diffusion algorithm to broadcast the area value of the current node to all directly connected neighboring nodes. After receiving the area data of neighboring nodes, the system calculates the aggregate area difference (absolute value) between the current node and each neighboring node. Then, this difference and the pre-stored edge length values ​​(representing the physical distance between nodes) in the topology graph are input into the weighted fusion engine. This engine multiplies the area difference by a negative weight coefficient (reflecting the weakening effect of difference on connection) and the edge length value by a positive weight coefficient (reflecting the weakening effect of distance on connection). The two are added together and then normalized by the Sigmoid function, finally outputting the standardized edge connection strength value (range 0-1). The larger this value, the better the continuity between the two aggregate regions; conversely, it indicates the existence of material distribution discontinuity.

[0078] Next, in step 1043, the system traverses all edges in the distribution topology graph, sequentially reads their connection strength values, and compares them with a preset threshold (a lower limit of connection strength set according to the aggregate gradation standard, which can be 0.3). When the connection strength value of an edge is lower than this threshold, the system triggers a boundary marking protocol, marking the edge as a segregation boundary (i.e., an abnormal connection where the aggregate distribution is broken or thinned). This process relies on graph traversal algorithms (such as depth-first search) to ensure that all edges are detected, and adapts to the process requirements of different mortar mix proportions through a dynamic threshold determination mechanism, ultimately generating a set of markers containing all abnormal edges.

[0079] Finally, in step 1044, the system performs boundary closure operations on all edges marked as segregation boundaries: based on graph theory connected component analysis, spatially adjacent segregation boundary edges are connected sequentially by their endpoint coordinates to form continuous paths (e.g., filling boundary gaps through Delaunay triangulation). Subsequently, a contour tracing algorithm (e.g., Moore neighborhood tracing) is used to transform the continuous paths into closed boundary contours. The physical region corresponding to the node enclosed by this contour is defined as the closed boundary of the segregation region, whose internal aggregate content is significantly lower than that of the surrounding region. The system outputs the set of polygon vertex coordinates of this closed boundary to guide the adjustment of mixing uniformity or defect repair.

[0080] In practical applications, in the aggregate segregation monitoring system of high-strength mortar production line, the system is based on the distribution topology map of high-strength mortar materials on the conveyor belt (a network structure composed of aggregate aggregation nodes and edge length values), and extracts the position coordinates of each node (e.g., the coordinates of node A are X=120cm, Y=80cm) and the actual area of ​​the aggregate area corresponding to the node (the actual pixel area of ​​the aggregate cluster is calculated by infrared scanning and image segmentation algorithm and converted into physical area).

[0081] Subsequently, the system transmits the actual area of ​​the aggregate region to adjacent nodes starting from each node (e.g., node A transmits its actual area of ​​the aggregate region to adjacent node B), calculates the difference between the actual areas of the aggregate regions of the node and the adjacent nodes (the absolute difference between the areas of node A and node B), and performs a weighted fusion calculation with the difference (representing the difference in aggregate density) and the edge length value (representing the physical distance between nodes) (the weighting coefficient is calibrated according to the mortar formula, and the fusion formula is: connection strength = edge length value / (1 + area difference)) to obtain the connection strength value of the edge.

[0082] After traversing the connection strength values ​​of all edges in the distribution topology graph, when the connection strength value is lower than the preset threshold (the threshold is set according to the mortar segregation sensitivity), the system marks the edge corresponding to the connection strength value as a segregation boundary that represents the abnormal distribution of aggregate (such as the edge of node AB being marked because the connection strength is too low).

[0083] Finally, the system connects all the edges marked as segregation boundaries (connecting adjacent segregation boundary edges using a graph traversal algorithm) to form a closed boundary contour with continuous paths (such as a rectangular contour composed of edges AB, BC, CD, and DA). The area of ​​nodes enclosed by the closed boundary contour constitutes the closed boundary of the segregation region (the actual area of ​​aggregates within this closed boundary varies greatly and the node spacing is abnormal, which is determined to be a segregation defect area). The system then drives the mixer to automatically compensate for cementitious materials or adjust the aggregate ratio to ensure that the mortar meets the homogeneity requirements for aggregate distribution.

[0084] The overall solution in step 104 above achieves intelligent identification and region segmentation of material segregation boundaries. Through connection strength analysis of the distribution topology map, an innovative topology feature transfer algorithm is designed to accurately identify weak connection edges characterizing aggregate distribution anomalies and automatically generate closed segregation region boundaries. This technology overcomes the limitations of traditional threshold segmentation, realizing intelligent segregation detection based on spatial relationship reasoning, and providing a precise basis for region division in quality assessment.

[0085] 105. Calculate the area ratio and spacing dispersion of the aggregate region within the closed boundary to generate the quantitative characteristics of the segregation region. At the same time, obtain the material movement speed and trigger the coordinated adjustment of the stirring speed and the proportion of cementitious materials based on the mapping relationship between the quantitative characteristics and the material movement speed.

[0086] Optionally, step 105 may specifically include the following steps: 1051. Within the closed boundary of the segregation region, the total area of ​​all aggregate regions is calculated, and the area ratio of the aggregate regions is obtained based on the total area and the total area of ​​the closed boundary.

[0087] 1052. Extract the average value of the side length between all adjacent aggregate region nodes within the closed boundary, and calculate the deviation coefficient of the side length value from the average value as the spacing dispersion.

[0088] 1053. Install a shaft speed sensor on the rear of the conveyor belt to obtain the current material movement speed in real time, and combine the area ratio value with the spacing dispersion into a quantitative feature structure.

[0089] 1054. Based on the preset mapping relationship between the material movement speed and the quantitative feature structure, determine the target value offset of the current stirring speed and the component increment value of the cementitious material ratio.

[0090] 1055. The target value offset is converted into a frequency control command for the stirring motor, and the component increment value is converted into an opening control command for the feeding valve to trigger the coordinated adjustment action of the execution equipment.

[0091] In the above scheme, area ratio refers to the ratio of the aggregate area to the total area of ​​the closed boundary. Spacing dispersion refers to the degree of dispersion in the spacing between adjacent aggregate areas. Quantitative feature refers to the numerical characteristics used to describe the segregation zone. Material movement speed refers to the speed at which the material moves on the conveyor belt. Mapping relationship refers to the correspondence between quantitative features and movement speed. Mixing speed refers to the rotational speed of the mixing equipment. Cementitious material ratio refers to the proportion of cementitious material in dry mortar. Total area value refers to the sum of the areas of all aggregate areas within the closed boundary. Total area of ​​the closed boundary refers to the total area of ​​the segregation zone. Deviation coefficient refers to the degree to which the spacing data deviates from the average value. Quantitative feature structure refers to the data structure containing area ratio and spacing dispersion. Target value offset refers to the value that the mixing speed needs to be adjusted. Component increment value refers to the proportion that the cementitious material needs to be increased by. Frequency control command refers to the control signal that adjusts the motor speed. Opening control command refers to the control signal that adjusts the valve opening degree. Coordinated adjustment action refers to the operation of multiple devices adjusting synchronously. Shaft speed sensor refers to the sensing device that measures the rotational speed. Feed valve refers to the valve device that controls the material supply. A stirring motor is an electric motor that drives stirring equipment.

[0092] In this embodiment, firstly, through step 1051, the system calculates the total area of ​​the closed boundary of the segregated region using a geometric area calculation engine. Then, the system traverses all aggregate regions within the boundary (i.e., regions marked as nodes in the distribution topology map), accumulating the actual area values ​​of each aggregate region (aggregate region attributes obtained in step 1041) to generate the total area value of all aggregate regions. Finally, the system divides the total area value by the total area of ​​the closed boundary, outputting the area ratio of the aggregate regions. This value reflects the enrichment degree of aggregates within the segregated region; a value lower than the standard value indicates excessive cementitious material.

[0093] Subsequently, in step 1052, the system extracts the edge length values ​​between all adjacent aggregate region nodes within the closed boundary (calculated in step 1034). The statistical characteristics of these edge length data are calculated using a dispersion analysis algorithm: first, the average value of all edge lengths is calculated; then, the absolute difference between each edge length value and the average value is calculated; finally, all differences are averaged and divided by the average value to generate a standardized spacing dispersion coefficient. This coefficient characterizes the uniformity of aggregate distribution; a higher coefficient indicates greater fluctuation in aggregate spacing and more uneven distribution, which can easily lead to uneven concrete strength.

[0094] Next, in step 1053, the system installs a speed sensor (such as a Hall effect encoder) on the shaft at the rear end of the conveyor belt. The system monitors the shaft rotation frequency in real time through pulse counting and converts it into the material movement speed. Simultaneously, the system combines the area percentage value and the spacing dispersion into a quantified feature structure—a data object containing two fields—used to encapsulate the quantified indicators of the separated regions.

[0095] Then, in step 1054, the system calls the pre-stored mapping relationship database (a regression model trained based on historical process data) and queries the matching quantitative feature structure benchmark value according to the current material movement speed. If the deviation between the measured structure and the benchmark value exceeds the allowable range, the system calculates the target value offset (the percentage by which the stirring speed needs to be increased) and the component increment value (the percentage by which the cementitious material needs to be added). For example, when the area ratio is lower than the benchmark and the spacing dispersion is too high, it triggers the addition of cementitious material to fill the aggregate gaps, while reducing the stirring speed to avoid material splashing.

[0096] Finally, in step 1055, the system inputs the target value offset into the frequency converter to generate a frequency control command for the stirring motor (e.g., the frequency can be adjusted from 50Hz to 55Hz); simultaneously, it converts the component increment value into an opening control command for the feeding valve (e.g., the opening can be increased from 60% to 70%). The commands are synchronously sent to the stirring motor and feeding valve via the industrial bus, triggering coordinated adjustment actions from the execution equipment: the motor speeds up to enhance material mixing intensity, and the valve expands to replenish the cementing material, jointly correcting segregation defects.

[0097] In practical applications, in the real-time control system for segregation in thermal insulation mortar production lines, during the continuous production of thermal insulation mortar, the system, based on the closed boundaries of the identified segregation areas on the conveyor belt (closed boundary contours generated by topological features), calculates the total area of ​​all aggregate areas (the physical area is calculated cumulatively by aggregate pixel clusters under high-contrast optical features). Based on the total area and the total area of ​​the closed boundary (the closed boundary is calibrated in real time by an infrared contour scanner), the system obtains the area ratio of the aggregate areas (characterizing the degree of aggregate enrichment within the segregation area). Simultaneously, it extracts the average side length values ​​between all adjacent aggregate area nodes within the closed boundary (nodes are aggregate aggregation centers marked by the distribution topology map), and calculates the deviation coefficient of the side length values ​​from the average value as the spacing dispersion (this coefficient quantifies the dispersion of aggregate distribution through the standard deviation algorithm; high dispersion indicates severe separation between aggregate and cementitious materials).

[0098] Simultaneously, the system installs a shaft speed sensor (such as an encoder or Hall sensor) on the conveyor belt to acquire the current material movement speed (conveyor belt linear speed) in real time. It combines the area ratio value and the spacing dispersion into a quantified feature structure (the structure contains two dimensions: area ratio and spacing dispersion). Then, based on the preset mapping relationship between the material movement speed and the quantified feature structure (the mapping relationship is generated based on historical discrete data training; for example, when the material movement speed is high, the area ratio exceeds the limit and the stirring speed needs to be reduced), it determines the target value offset of the current stirring speed and the component increment value of the cementitious material ratio (for example, when the area ratio is too high and the spacing dispersion is large, the target value offset is to reduce the stirring motor speed, and the component increment value is to increase the cementitious material feeding ratio).

[0099] Ultimately, the system converts the target value offset into a frequency control command for the mixing motor (adjusting the motor speed via a frequency converter), and simultaneously converts the component increment value into an opening control command for the feeding valve (controlling the opening of the electric valve in the cementitious material silo). This triggers coordinated adjustment actions from the execution equipment (the mixer decelerates to extend the mixing time, and the cementitious material increment compensates for the lack of bonding in the aggregate enrichment area), suppressing segregation defects in real time and ensuring that the thermal insulation mortar meets the process requirements for the homogeneity of aggregate distribution.

[0100] The overall solution in step 105 above achieves quantitative analysis of segregation characteristics and intelligent adjustment of production parameters. By calculating quantitative indicators such as the area ratio and spacing dispersion of the segregation region, and combining this with real-time monitoring of material movement speed, an innovative mapping relationship model for production parameter adjustment is constructed. This technology can automatically generate coordinated control commands for stirring speed and cementitious material ratio, realizing closed-loop management from quality inspection to production control, and significantly improving the quality control level of dry mortar production.

[0101] The following are specific examples of steps 101 to 105, such as Figure 2 As shown: In the continuous production process of crack-resistant mortar, the system first projects a multi-angle incident light field (including visible light and near-infrared bands) onto the surface of the material moving at a uniform speed on the conveyor belt through a ring light source array. Simultaneously, it collects reflected light data, extracts the reflection intensity values ​​of aggregate and cementitious material, and confirms that the difference in spectral reflection characteristics is that the aggregate has a higher reflection intensity in a specific band. Based on this difference, the incident light wavelength and irradiation angle of the ring light source array are adjusted in real time (such as enhancing the irradiation angle of the near-infrared band). After reprojection, the reflection intensity of the aggregate edge is significantly higher than that of the cementitious material area, forming a high-contrast optical feature at the aggregate edge.

[0102] Subsequently, the optical sensor collects the reflected light under this feature to obtain multispectral reflectance data containing multiple spectral bands of reflectance intensity values; based on the characteristic that the reflectance intensity value of the aggregate is always higher than that of the cementitious material, the continuous spectral band range that meets this characteristic is selected and determined as the characteristic band corresponding to the aggregate; this characteristic band is combined with the range outside the characteristic band to form a reflectance structure, which is defined as the reflectance characteristic vector of the aggregate and cementitious material.

[0103] Next, the system acquires the reflection feature vectors of all observation units and extracts the average component of the reflection intensity value of the aggregate (characterizing the aggregate content level). The system triggers the synchronous pulse signal of the optical sensor by using positioning markers on both sides of the conveyor belt, and combines this with the counting value of the rotary encoder to determine the position coordinates of each observation unit. This establishes a mapping relationship between the position coordinates and the aggregate content level, forming the spatial distribution relationship of the reflection feature vectors. Continuous observation unit groups with content levels exceeding a threshold are extracted, and their center positions are used as nodes in the distribution topology map (marking the actual distribution location of the aggregate area). The straight-line distance between adjacent nodes is calculated as the edge length value, and finally, the actual distribution location and the edge length value are combined to generate the distribution topology map.

[0104] Subsequently, the system extracts the location coordinates of each node and the actual area of ​​the aggregate region from the distribution topology map; it transfers the actual area from the node to the adjacent node, calculates the difference between the actual area of ​​the aggregate region of the node and the adjacent node, and calculates the connection strength value of the edge by weighted fusion with the edge length value; after traversing all edges, the edges with connection strength values ​​lower than the preset threshold are marked as segregation boundaries; and all segregation boundaries are connected into a closed boundary contour through connection operations, and the area enclosed by it constitutes the closed boundary of the segregation region.

[0105] Finally, the total area of ​​all aggregate regions within the closed boundary is calculated, and the area ratio of the aggregate regions is calculated based on the total area of ​​the closed boundary. The average side length of all adjacent nodes within the closed boundary is extracted, and the deviation coefficient of the side length value from the average value is calculated as the spacing dispersion. At the same time, the material movement speed is obtained through the shaft speed sensor, and the area ratio and spacing dispersion are combined into a quantitative feature structure. According to the preset mapping relationship between the material movement speed and the quantitative feature structure, the target value offset of the stirring speed and the component increment value of the cementitious material ratio are determined. The target value offset is converted into a frequency control command for the stirring motor, and the component increment value is converted into an opening control command for the feeding valve, triggering coordinated adjustment actions (such as reducing the stirring speed and increasing the cementitious material feeding ratio) to suppress segregation defects in real time.

[0106] Figure 3 This application provides a schematic diagram of the structure of a real-time defect identification system for dry mortar production, as shown in the embodiment of the present application. Figure 3 As shown, the system includes: The adjustment module 31 is used to project a multi-angle incident light field onto the surface of the continuously moving dry powder mortar through a ring light source array, and to obtain the difference in spectral reflectance characteristics between the aggregate and cementitious material in the material. Based on the difference in spectral reflectance characteristics, the incident wavelength and irradiation angle of the multi-angle incident light field are adjusted to form a high-contrast optical feature at the edge of the aggregate. The separation module 32 is used to collect multispectral reflectance data under the high-contrast optical features, and based on the characteristic that the reflectance intensity value of the aggregate is higher than that of the cementitious material, to separate the characteristic band corresponding to the aggregate from the multispectral reflectance data to generate the reflectance feature vectors of the aggregate and the cementitious material. The acquisition module 33 is used to acquire the spatial distribution relationship of the reflection feature vector and construct the distribution topology map of aggregate on the conveyor belt based on the spatial distribution relationship. The nodes in the distribution topology map represent aggregate regions, and the edges in the distribution topology map represent the spacing between adjacent aggregate regions. The update module 34 is used to transfer topological structure features of the distribution topology map to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a segregated region, and the closed boundary of the segregated region is segmented. The trigger module 35 is used to calculate the area ratio and spacing dispersion of the aggregate region within the closed boundary to generate the quantitative characteristics of the segregation region. At the same time, it acquires the material movement speed and triggers the coordinated adjustment of the stirring speed and the proportion of cementitious materials based on the mapping relationship between the quantitative characteristics and the material movement speed.

[0107] Figure 3 The aforementioned real-time defect identification system for dry mortar production can perform... Figure 1 The implementation principle and technical effects of the real-time identification method for dry mortar production defects described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the real-time identification system for dry mortar production defects in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here. In one possible design, Figure 3 The real-time defect identification system for dry mortar production shown in the embodiment can be implemented as a computing device, such as... Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42; The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 42.

[0108] The processing component 42 is used for the above Figure 1 The embodiment describes a method for real-time identification of defects in dry mortar production.

[0109] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0110] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0111] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0112] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0113] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0114] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0115] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a method for real-time identification of defects in dry mortar production.

[0116] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0117] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0118] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, 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 cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions 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 this application.

Claims

1. A method for real-time identification of defects in dry powder mortar production, characterized in that, include: A multi-angle incident light field is projected onto the surface of a continuously moving dry powder mortar material by a ring light source array, and the difference in spectral reflectance characteristics between aggregates and cementitious materials in the material is obtained. The incident wavelength and irradiation angle of the multi-angle incident light field are adjusted according to the difference in spectral reflectance characteristics to form high-contrast optical features at the edge of the aggregate. Multispectral reflectance data under the high-contrast optical characteristics are collected, and based on the characteristic that the reflectance intensity value of aggregate is higher than that of cementitious material, the characteristic bands corresponding to aggregate are separated from the multispectral reflectance data to generate the reflectance feature vectors of aggregate and cementitious material. The spatial distribution relationship of the reflection feature vectors is obtained, and a distribution topology map of aggregates on the conveyor belt is constructed based on the spatial distribution relationship. Nodes in the distribution topology map represent aggregate regions, and edges in the distribution topology map represent the spacing between adjacent aggregate regions, including: Obtain the reflection feature vectors of all observation units on the conveyor belt, and extract the average component of the reflection intensity value of the aggregate in the reflection feature vectors. The average component represents the content level of aggregate in each observation unit. The position coordinates of the observation unit on the conveyor belt are located, and a mapping relationship between the position coordinates and the aggregate content level is established. The mapping relationship constitutes the spatial distribution relationship of the reflection feature vector. Extract continuous observation unit groups where the aggregate content level on the conveyor belt exceeds a preset threshold, and use the center position of each continuous observation unit group as a node in the distribution topology map, where the node represents the actual distribution position of the aggregate area; Calculate the straight-line distance between any two adjacent aggregate regions and use the straight-line distance as the edge length value connecting the two nodes in the distribution topology graph; The actual distribution locations and side length values ​​are combined into a mesh structure to serve as a distribution topology map characterizing the distribution of aggregates on the conveyor belt; The topological structure features of the distributed topology graph are transferred to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary, and the closed boundary of the separation region is segmented, including: Extract the location coordinates of each node and the actual area of ​​the aggregate region corresponding to the node from the distribution topology map; Starting from each node, the actual area of ​​the aggregate region is transmitted to the adjacent nodes, and the difference between the actual area of ​​the aggregate region of the node and the adjacent nodes is calculated. The difference is then weighted and fused with the edge length value to obtain the connection strength value of the edge. Traverse the connection strength values ​​of all edges in the distribution topology graph. When the connection strength value is lower than a preset threshold, mark the edge corresponding to the connection strength value as a segregation boundary that characterizes the abnormal distribution of aggregate. Connect all edges marked as separation boundaries to form a closed boundary contour with a continuous path, and the region of the nodes enclosed by the closed boundary contour constitutes the closed boundary of the separation region. Calculate the area ratio and spacing dispersion of the aggregate region within the closed boundary to generate the quantitative characteristics of the segregation region. Simultaneously, acquire the material movement velocity and, based on the mapping relationship between the quantitative characteristics and the material movement velocity, trigger the coordinated adjustment of the stirring speed and the proportion of cementitious materials, including: Within the closed boundary of the segregation region, the total area of ​​all aggregate regions is calculated, and the area ratio of the aggregate regions is obtained based on the total area and the total area of ​​the closed boundary. Extract the average value of the edge length between all adjacent aggregate region nodes within the closed boundary, and calculate the deviation coefficient of the edge length value from the average value as the spacing dispersion; A speed sensor is installed at the rear shaft of the conveyor belt. The rotation frequency of the conveyor belt shaft is detected in real time by pulse counting. The linear speed of the conveyor belt is calculated based on the rotation frequency. The linear speed of the conveyor belt is used as the current material movement speed. At the same time, the area ratio value and the spacing dispersion are combined into a quantitative feature structure. Based on the preset mapping relationship between the material movement speed and the quantified feature structure, the target value offset of the current stirring speed and the component increment value of the cementitious material ratio are determined. The target value offset is converted into a frequency control command for the stirring motor, and the component increment value is converted into an opening control command for the feeding valve, so as to trigger the coordinated adjustment action of the execution equipment.

2. The method according to claim 1, characterized in that, A multi-angle incident light field is projected onto the surface of continuously moving dry powder mortar using a ring light source array, and the differences in spectral reflectance characteristics between aggregates and cementitious materials are obtained. Based on these differences, the incident wavelength and illumination angle of the multi-angle incident light field are adjusted to create high-contrast optical features at the aggregate edges. This includes: By simultaneously projecting incident light fields at different irradiation angles onto the surface of continuously moving dry powder mortar material through a ring light source array, the reflected light data of the material surface is obtained at the same time. Extract the aggregate reflection intensity value and the cementitious material reflection intensity value from the reflected light data, and extract the difference in spectral reflection characteristics between the aggregate and the cementitious material from the reflected light data. The difference in spectral reflection characteristics is that the reflection intensity value of the aggregate is higher than that of the cementitious material. Based on the differences in spectral reflectance characteristics, the incident light wavelength of the ring light source array and the illumination angle of each angle light source are adjusted in real time. The incident light wavelength and irradiation angle are adjusted and re-projected onto the material surface, so that the intensity of reflected light in the aggregate edge area is significantly higher than that in the cementitious material area, thus forming a high-contrast optical feature at the aggregate edge.

3. The method according to claim 1, characterized in that, Locate the position coordinates of the observation unit on the conveyor belt, and establish a mapping relationship between the position coordinates and the aggregate content level. This mapping relationship constitutes the spatial distribution relationship of the reflection feature vector, including: Positioning markers are set at equal intervals on both sides of the conveyor belt. When the observation unit moves with the conveyor belt and passes the positioning markers, it triggers the synchronization pulse signal of the optical sensor. The position coordinates of each observation unit in the conveyor belt are determined by matching the timestamp of the synchronization pulse signal with the count value of the rotary encoder of the conveyor belt. A mapping relationship is established between the actual location coordinates of each observation unit and the content level of the aggregate to form a data unit that combines location and content. Collect data units of location and content combinations of all observation units, and construct a spatial distribution relationship describing the aggregate distribution state on the material surface according to the location coordinates.

4. The method according to claim 1, characterized in that, Multispectral reflectance data under the high-contrast optical characteristics are collected, and based on the characteristic that the reflectance intensity of aggregate is higher than that of cementitious material, the characteristic bands corresponding to the aggregate are separated from the multispectral reflectance data to generate reflectance feature vectors of aggregate and cementitious material, including: The reflected light from the surface of a material under high-contrast optical characteristics is collected by an optical sensor to obtain multispectral reflectance data containing multiple spectral band reflectance intensity values; Based on the characteristic that the reflection intensity value of aggregate is always higher than that of cementitious material, a continuous spectral band range that meets the above characteristic is selected from the complete spectral bands of multispectral reflectance data, and the continuous spectral band range is determined as the characteristic band corresponding to the aggregate. The characteristic band of the aggregate and the range outside the characteristic band are combined to form a reflection structure, and the reflection structure is defined as the reflection characteristic vector of the aggregate and cementitious material.

5. A real-time defect identification system for dry powder mortar production, characterized in that, include: The adjustment module is used to project a multi-angle incident light field onto the surface of a continuously moving dry powder mortar material through a ring light source array, and to obtain the difference in spectral reflectance characteristics between aggregates and cementitious materials in the material. Based on the difference in spectral reflectance characteristics, the incident wavelength and irradiation angle of the multi-angle incident light field are adjusted to form a high-contrast optical feature at the edge of the aggregate. The separation module is used to collect multispectral reflectance data under the high-contrast optical features, and based on the characteristic that the reflectance intensity value of the aggregate is higher than that of the cementitious material, it separates the characteristic band corresponding to the aggregate from the multispectral reflectance data to generate the reflectance feature vectors of the aggregate and the cementitious material. The acquisition module is used to acquire the spatial distribution relationship of the reflection feature vectors and construct a distribution topology map of aggregates on the conveyor belt based on the spatial distribution relationship. Nodes in the distribution topology map represent aggregate regions, and edges in the distribution topology map represent the spacing between adjacent aggregate regions. This includes: Obtain the reflection feature vectors of all observation units on the conveyor belt, and extract the average component of the reflection intensity value of the aggregate in the reflection feature vectors. The average component represents the content level of aggregate in each observation unit. The position coordinates of the observation unit on the conveyor belt are located, and a mapping relationship between the position coordinates and the aggregate content level is established. The mapping relationship constitutes the spatial distribution relationship of the reflection feature vector. Extract continuous observation unit groups where the aggregate content level on the conveyor belt exceeds a preset threshold, and use the center position of each continuous observation unit group as a node in the distribution topology map, where the node represents the actual distribution position of the aggregate area; Calculate the straight-line distance between any two adjacent aggregate regions and use the straight-line distance as the edge length value connecting the two nodes in the distribution topology graph; The actual distribution locations and side length values ​​are combined into a mesh structure to serve as a distribution topology map characterizing the distribution of aggregates on the conveyor belt; The update module is used to transfer topological structure features to the distribution topology map to update the connection strength between nodes. When the connection strength is lower than a preset threshold, it is determined to be a separation boundary, and the closed boundary of the separation region is segmented, including: Extract the location coordinates of each node and the actual area of ​​the aggregate region corresponding to the node from the distribution topology map; Starting from each node, the actual area of ​​the aggregate region is transmitted to the adjacent nodes, and the difference between the actual area of ​​the aggregate region of the node and the adjacent nodes is calculated. The difference is then weighted and fused with the edge length value to obtain the connection strength value of the edge. Traverse the connection strength values ​​of all edges in the distribution topology graph. When the connection strength value is lower than a preset threshold, mark the edge corresponding to the connection strength value as a segregation boundary that characterizes the abnormal distribution of aggregate. Connect all edges marked as separation boundaries to form a closed boundary contour with a continuous path, and the region of the nodes enclosed by the closed boundary contour constitutes the closed boundary of the separation region. A triggering module is used to calculate the area ratio and spacing dispersion of the aggregate region within the closed boundary to generate the quantitative characteristics of the segregation region. Simultaneously, it acquires the material movement speed and, based on the mapping relationship between the quantitative characteristics and the material movement speed, triggers the coordinated adjustment of the stirring speed and the proportion of cementitious materials, including: Within the closed boundary of the segregation region, the total area of ​​all aggregate regions is calculated, and the area ratio of the aggregate regions is obtained based on the total area and the total area of ​​the closed boundary. Extract the average value of the edge length between all adjacent aggregate region nodes within the closed boundary, and calculate the deviation coefficient of the edge length value from the average value as the spacing dispersion; A speed sensor is installed at the rear shaft of the conveyor belt. The rotation frequency of the conveyor belt shaft is detected in real time by pulse counting. The linear speed of the conveyor belt is calculated based on the rotation frequency. The linear speed of the conveyor belt is used as the current material movement speed. At the same time, the area ratio value and the spacing dispersion are combined into a quantitative feature structure. Based on the preset mapping relationship between the material movement speed and the quantified feature structure, the target value offset of the current stirring speed and the component increment value of the cementitious material ratio are determined. The target value offset is converted into a frequency control command for the stirring motor, and the component increment value is converted into an opening control command for the feeding valve, so as to trigger the coordinated adjustment action of the execution equipment.

6. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a real-time identification method for dry mortar production defects as described in any one of claims 1 to 4.

7. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements a real-time identification method for dry mortar production defects as described in any one of claims 1 to 4.

Images