Industrial salt crystalline particle size distribution on-line detection method and system based on machine vision

Abstract

The application provides an industrial salt crystallization particle size distribution online detection method and system based on machine vision. In the application, the trajectory tracking of salt crystallization particles is performed on the obtained dynamic image sequence to obtain a motion trajectory set, according to which the spatial distribution characteristics of the salt crystallization particles in the saturated solution are determined, the spatial distribution characteristics are corrected by combining the real-time temperature parameters and the real-time concentration parameters, the particle distribution information is obtained, the initial particle size distribution curve is generated and adjusted based on the particle distribution information, the target particle size distribution curve is obtained, the matching degree analysis of the target particle size distribution curve and the preset industrial salt crystallization quality standard is performed, and the online detection result is output. The technical scheme provided by the application not only overcomes the inherent defects of the traditional offline sampling method which is separated from the actual production conditions, but also provides a reliable data basis for the instant analysis and accurate control of the production process.

Description

Technical Field

[0001] This application relates to the field of image analysis technology, and in particular to an online detection method and system for the particle size distribution of industrial salt crystals based on machine vision. Background Technology

[0002] In the evaporation and crystallization production of industrial salt, real-time and accurate detection of the particle size distribution of crystallized particles is crucial to ensuring the purity and uniformity of the final product and optimizing the crystallization process. The crystallization process occurs in a dynamic and non-uniform saturated solution environment, where particles continuously move, collide, and grow. Traditional offline sampling and detection methods suffer from severe lag and cannot capture the transient changes in the process.

[0003] Existing technologies automatically collect small amounts of salt slurry samples at specific points in the crystallization process, place them on a stationary stage, and capture clear static images using an industrial camera. Subsequently, digital image processing technology is used to identify and measure the size of the salt crystal particles in the sample, thereby calculating the particle size distribution of the batch of samples.

[0004] However, this approach has inherent flaws. It is essentially an offline, static sampling and testing method. The analyzed samples have been removed from the real, flowing environment inside the crystallizer. The movement, spatial distribution, and interaction of the particles are artificially stopped, resulting in measurement results that cannot truly reflect the entire continuous and dynamic crystallization process on the production line. This method completely ignores the direct impact of real-time operating parameters on particle behavior and test results. The obtained particle size distribution data is isolated and fails to be coupled with the process status, thus it cannot be used to achieve true real-time closed-loop process optimization and control. Summary of the Invention

[0005] This application provides a machine vision-based online detection method and system for industrial salt crystallization particle size distribution, which solves the problems in the prior art where offline and static sampling and analysis methods fail to accurately reflect the entire continuous and dynamic crystallization process, and where particle size distribution data is isolated and cannot be used for real-time process optimization and control because the direct impact of real-time operating parameters on particle behavior and detection results is completely ignored.

[0006] In a first aspect, this application provides an online detection method for the particle size distribution of industrial salt crystals based on machine vision, including:

[0007] Acquire dynamic image sequences during the evaporation and crystallization process of industrial salt;

[0008] The motion trajectory set is obtained by tracing the trajectory of salt crystal particles in the dynamic image sequence.

[0009] Based on the set of motion trajectories, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined;

[0010] By combining real-time temperature and concentration parameters during the industrial salt evaporation and crystallization process, the spatial distribution characteristics are corrected using multi-physics field coupling to obtain particle distribution information.

[0011] An initial particle size distribution curve for industrial salt crystallization is generated based on the particle distribution information, and the initial particle size distribution curve is adjusted to obtain the target particle size distribution curve.

[0012] The target particle size distribution curve is matched with the preset industrial salt crystallization quality standard, and the online detection results of the industrial salt crystallization particle size distribution are output.

[0013] Optionally, the dynamic image sequence is subjected to trajectory tracking of salt crystal particles to obtain a set of motion trajectories, including:

[0014] Identify the position coordinates of salt crystal particles within each image of the dynamic image sequence;

[0015] The position coordinates of salt crystal particles in two adjacent images in the dynamic image sequence are correlated to determine the positional changes of the salt crystal particles;

[0016] The positional changes of the salt crystal particles on multiple consecutive dynamic images are connected to form an initial motion trajectory;

[0017] The initial motion trajectory was identified, and the target trajectory was found to be abnormal.

[0018] The abnormality of the target trajectory is corrected, and the corrected initial motion trajectory is integrated with the initial motion trajectory that is not affected by the abnormality of the target trajectory to form a motion trajectory set.

[0019] Optionally, based on the set of motion trajectories, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined, including:

[0020] Based on the set of motion trajectories, the three-dimensional spatial position sequence of salt crystal particles in the crystallization tank is extracted;

[0021] Based on the three-dimensional spatial position sequence, calculate the distribution density of salt crystal particles in the axial and radial directions of the crystallization tank;

[0022] Based on the three-dimensional spatial position sequence, the movement direction of salt crystal particles in the saturated solution is analyzed, and salt crystal particles with the same movement direction are divided into the same movement group.

[0023] Based on the distribution density and the motion group, a high-density motion zone and a low-density motion zone are divided within the crystallization tank;

[0024] Based on the spatial relationship between the high-density motion region and the low-density motion region, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined.

[0025] Optionally, by combining real-time temperature and concentration parameters during the industrial salt evaporation and crystallization process, the spatial distribution characteristics are corrected using multi-physics coupling to obtain particle distribution information, including:

[0026] Based on the high-density and low-density motion regions in the spatial distribution characteristics, an initial particle distribution model is established within the crystallization tank.

[0027] The settling velocity of salt crystal particles in the initial particle distribution model is corrected based on real-time temperature parameters to obtain the corrected settling velocity distribution.

[0028] The aggregation state of salt crystal particles in the initial particle distribution model is corrected based on the real-time concentration parameters to obtain the corrected aggregation state distribution.

[0029] By fusing the corrected sedimentation velocity distribution and the corrected aggregation state distribution, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours.

[0030] The updated boundary contours are integrated into the initial particle distribution model to output particle distribution information through the integrated particle distribution model.

[0031] Optionally, by fusing the corrected settlement velocity distribution and the corrected aggregation state distribution, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours, including:

[0032] Establish the correspondence between the corrected settlement velocity distribution and the corrected aggregation state distribution;

[0033] Based on the aforementioned correspondence, the combined influence parameters on the spatial distribution of particles when the modified settling velocity distribution and the modified aggregation state distribution work together are determined.

[0034] Based on the composite influence parameters, calculate the movement vector of each point on the boundary between the high-density motion region and the low-density motion region;

[0035] Adjust the spatial position of each point on the boundary according to the movement vector to form a boundary point set;

[0036] Based on the set of boundary points, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours.

[0037] Optionally, an initial particle size distribution curve for industrial salt crystallization is generated based on the particle distribution information, and the initial particle size distribution curve is adjusted to obtain a target particle size distribution curve, including:

[0038] Based on the spatial location and particle size characteristics of the salt crystal particles in the particle distribution information, the number of salt crystal particles in different particle size ranges is counted to generate an initial particle size distribution curve.

[0039] Extract the differences in particle size characteristics between salt crystal particles in high-density motion regions and low-density motion regions from the particle distribution information;

[0040] Based on the differences in particle size characteristics, calculate the adjustment parameters for different particle size segments in the initial particle size distribution curve;

[0041] The adjustment parameters are applied to the corresponding particle size segment of the initial particle size distribution curve to form an intermediate particle size distribution curve;

[0042] Based on the spatial uniformity of salt crystal particles in the particle distribution information, the intermediate particle size distribution curve is smoothed to generate the target particle size distribution curve.

[0043] Optionally, the target particle size distribution curve is compared with a preset industrial salt crystallization quality standard to perform a matching degree analysis, and the online detection results of the industrial salt crystallization particle size distribution are output, including:

[0044] The target particle size distribution curve is decomposed into multiple target segments according to the particle size interval, and the preset numerical range of different particle size intervals in the preset industrial salt crystallization quality standard is decomposed into multiple standard segments.

[0045] Based on the target segment and the standard segment, the distribution value difference between each target segment and the corresponding standard segment is compared one by one to obtain the matching deviation of each granularity segment;

[0046] Based on the matching deviation of each particle size segment, a weighted comprehensive calculation is performed using the weight values ​​preset for different particle size segments in the industrial salt crystallization quality standard to obtain the comprehensive matching degree.

[0047] The overall matching degree is compared with a preset qualified threshold, and the quality grade of the industrial salt crystallization product is determined based on the comparison result.

[0048] The key feature parameters extracted from the target particle size distribution curve are combined with the quality level to generate online detection results.

[0049] Secondly, this application provides an online detection system for the particle size distribution of industrial salt crystals based on machine vision, comprising:

[0050] The acquisition module is used to acquire dynamic image sequences during the industrial salt evaporation and crystallization process;

[0051] The tracking module is used to track the trajectory of salt crystal particles in the dynamic image sequence to obtain a set of motion trajectories;

[0052] The determination module is used to determine the spatial distribution characteristics of salt crystal particles in a saturated solution based on the set of motion trajectories.

[0053] The correction module is used to combine real-time temperature parameters and real-time concentration parameters during the industrial salt evaporation and crystallization process to perform multi-physics field coupling correction on the spatial distribution characteristics to obtain particle distribution information.

[0054] The adjustment module is used to generate an initial particle size distribution curve for industrial salt crystallization based on the particle distribution information, and to adjust the initial particle size distribution curve to obtain the target particle size distribution curve.

[0055] The output module is used to perform a matching degree analysis between the target particle size distribution curve and the preset industrial salt crystallization quality standard, and output the online detection results of the industrial salt crystallization particle size distribution.

[0056] Thirdly, this application provides 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 the online detection method for industrial salt crystallization particle size distribution based on machine vision as described in the first aspect above.

[0057] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements an online detection method for industrial salt crystallization particle size distribution based on machine vision as described in the first aspect.

[0058] This application achieves continuous, online observation of salt crystallization particles in a real flow environment by directly acquiring dynamic image sequences and tracking particle trajectories during the industrial salt evaporation and crystallization process. This overcomes the inherent defects of traditional offline sampling methods that are detached from actual production conditions. By coupling and correcting the spatial distribution characteristics of particles obtained from visual analysis with real-time temperature and concentration parameters using multi-physics fields, the obtained particle distribution information and the final generated target particle size distribution curve can accurately reflect the real particle size distribution state under the combined effect of the dynamic crystallization process and real-time process conditions, providing a reliable data foundation for real-time analysis and precise control of the production process.

[0059] Furthermore, by conducting a refined matching degree analysis of the target particle size distribution curve and quality standards based on particle size segments, and introducing weights that conform to the importance of the production process for comprehensive evaluation, a multi-dimensional and quantitative evaluation of the quality of industrial salt crystallization products is achieved, avoiding the one-sidedness of single threshold judgment. The final online detection results not only contain clear quality grade conclusions, but also combine key characteristic parameters reflecting the distribution pattern, thereby providing production operators with complete decision-making information, significantly improving the precision of quality control and the practical guiding value of feedback results.

[0060] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in 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.

[0062] Figure 1 A flowchart of an online detection method for industrial salt crystal particle size distribution based on machine vision, as provided in this application, is shown.

[0063] Figure 2 A schematic diagram of the structure of an online detection system for industrial salt crystallization particle size distribution based on machine vision provided in this application is shown.

[0064] Figure 3 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0065] To enable those skilled in the art to better understand the present application, the technical solution of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0066] 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.

[0067] The technical solutions of this application will now be clearly and completely described 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.

[0068] Figure 1 This application provides a flowchart of an online detection method for the particle size distribution of industrial salt crystals based on machine vision, as shown below. Figure 1 As shown, the method includes:

[0069] Step 101: Obtain a dynamic image sequence of the industrial salt evaporation and crystallization process.

[0070] In the above steps, the industrial salt evaporation crystallization process refers to the continuous production process in which water in a sodium chloride aqueous solution is evaporated by heating in a crystallization tank, the solution concentration gradually increases and reaches a supersaturated state, thereby promoting the continuous formation, growth and precipitation of salt crystal particles from the solution.

[0071] Dynamic image sequences refer to the synchronous acquisition of images by multi-angle high-speed industrial cameras placed at the observation window of the crystallizer, used to record the position, shape, quantity and continuous movement of salt crystal particles in the solution within the crystallizer over time.

[0072] In this embodiment, firstly, multiple high-speed industrial cameras with high frame rate shooting capabilities are fixedly installed at the positions of transparent observation windows on both sides of the crystallization tank for industrial salt evaporation and crystallization, and the lens angle and focal length are adjusted to ensure that their field of view can cover the main crystallization reaction area inside the tank.

[0073] Secondly, during the industrial salt evaporation and crystallization process, a high-speed industrial camera is synchronously activated by the central controller to start image acquisition at the same and pre-set high frame rate. Each camera generates a digital image at each sampling moment.

[0074] Then, it continuously receives image data streams from high-speed industrial cameras, binds two images captured from different angles at the same millisecond time point into a group according to strict acquisition timestamps, and arranges and combines multiple groups of images captured at multiple consecutive times strictly according to the chronological order, thereby forming a dynamic image sequence.

[0075] For example, on the A-type evaporation and crystallization production line of a large industrial salt production enterprise, technicians installed high-speed industrial cameras of brand B outside the pressure-resistant glass observation windows on both sides of the crystallization tank. After the sodium chloride aqueous solution began to boil and crystallize, the operator started the image acquisition program in the central control room. The two cameras began to synchronously acquire images of the particulate suspension in the crystallization tank at a rate of 500 frames per second. The two images acquired from the left and right perspectives every 2 milliseconds were bound into a pair of images, stamped with a unified timestamp, and stored. After 10 seconds of continuous acquisition, a dynamic image sequence consisting of 5,000 sets of images that are strictly continuous in time and complementary in space was obtained.

[0076] This step successfully captured a dynamic image sequence by deploying a high-speed industrial camera and implementing synchronous high frame rate acquisition, thus visually capturing the dynamic and continuous industrial salt evaporation and crystallization process. This non-contact and uninterrupted recording of the real state of the production site fundamentally overcomes the shortcomings of traditional offline sampling methods, such as disrupting process continuity and having monitoring blind spots. It provides a unique, real, continuous, and high-quality data source for subsequent accurate particle trajectory tracking and behavior analysis, laying the data foundation for the entire online detection method.

[0077] Step 102: Track the trajectory of salt crystal particles in the dynamic image sequence to obtain a set of motion trajectories.

[0078] Optionally, step 102 may specifically include the following steps:

[0079] Step 1021: Identify the position coordinates of salt crystal particles in each image of the dynamic image sequence;

[0080] Step 1022: Correlate the position coordinates of salt crystal particles in two adjacent images in the dynamic image sequence to determine the positional changes of the salt crystal particles;

[0081] Step 1023: Connect the positional changes of the salt crystal particles on multiple consecutive dynamic images to form an initial motion trajectory;

[0082] Step 1024: Identify the initial motion trajectory to determine if the target trajectory is abnormal;

[0083] Step 1025: Correct the target trajectory anomaly, and integrate the corrected initial motion trajectory with the initial motion trajectory that is not affected by the target trajectory anomaly to form a motion trajectory set.

[0084] In the above steps, salt crystal particles refer to solid sodium chloride crystals that precipitate and grow from a supersaturated solution during the industrial salt evaporation and crystallization process. These are the core target objects for observation, tracking, and analysis in this invention.

[0085] Trajectory tracking refers to the technical process of analyzing a continuous sequence of images to identify the position of the same object in different images and connecting them in chronological order to depict the object's motion path.

[0086] The motion trajectory set refers to the set of information obtained by analyzing a sequence of dynamic images, which contains the complete motion path information of each of the successfully tracked salt crystal particles.

[0087] Position coordinates refer to the horizontal and vertical positions of the particle's center point within the two-dimensional plane of the image, obtained through image analysis techniques;

[0088] Position change refers to the change in the position coordinates of the same salt crystal particle between two adjacent images, used to describe the displacement direction and distance of the salt crystal particle within a very short time interval;

[0089] The initial motion trajectory refers to the preliminary motion path formed by simply connecting the position points of salt crystal particles in multiple consecutive frames of images in a time sequence, without considering the mutual interference between salt crystal particles in an ideal situation. This path may contain interruptions or errors.

[0090] Target trajectory anomaly refers to trajectory anomalies caused by salt crystal particles coming into contact with, overlapping or blocking each other during the initial motion trajectory.

[0091] The corrected initial trajectory refers to a more complete, coherent, and reasonable motion path obtained after processing an initial trajectory with target trajectory anomalies using a specific trajectory repair algorithm.

[0092] In this embodiment, each image in the dynamic image sequence is first processed using image segmentation technology in computer vision to identify the position coordinates of salt crystal particles in each image.

[0093] Secondly, through the data association step in the multi-target tracking algorithm, the position coordinates of salt crystal particles in two adjacent images are matched and associated to determine the position change of each salt crystal particle;

[0094] Next, through trajectory initialization and update algorithms, the position coordinates of the same salt crystal particle on multiple consecutive dynamic images are connected in chronological order to form the initial motion trajectory.

[0095] Then, by setting trajectory filtering and discrimination rules, all initial motion trajectories are scanned and analyzed to identify and mark trajectory anomalies with problems such as interruption, merging, or jump.

[0096] Finally, the trajectory repair algorithm is used to process the marked target trajectory anomalies, and these corrected initial motion trajectories are merged with those unaffected initial motion trajectories to form a motion trajectory set.

[0097] Following the previous implementation example, firstly, the image processing program reads these 10,000 images, applies image segmentation and target recognition algorithms to identify the outlines of all salt crystal particles in each image, and calculates the position coordinates of the salt crystal particles in each image.

[0098] Secondly, the data association module in the multi-target tracking algorithm is used to compare and match the position coordinates of all salt crystal particles in two adjacent images in the dynamic image sequence, thereby determining the position change of each salt crystal particle.

[0099] Next, the trajectory construction module connects the position coordinates of the same salt crystal particle, which are determined sequentially in multiple consecutive dynamic images, according to their time sequence to form the initial motion trajectory.

[0100] Afterwards, the trajectory analysis module scans all initial motion trajectories and identifies, according to preset rules, trajectory segments that are interrupted, merged, or abruptly changed due to contact or brief occlusion of salt crystal particles during movement. These marked trajectory segments are identified as trajectory anomalies.

[0101] Finally, the trajectory repair module is activated to correct the initial motion trajectory with trajectory anomalies and integrate it with the initial motion trajectory that is not affected by the trajectory anomaly to form a motion trajectory set.

[0102] This step transforms discrete images into continuous and reliable motion trajectory data through particle detection, cross-frame correlation, and trajectory construction, and corrects tracking errors caused by particle interactions. This dataset realistically reproduces the motion history and behavior patterns of particles in complex flow fields, overcoming the limitation of only observing static distributions, and providing a crucial dynamic data foundation for analyzing particle spatial characteristics, motion laws, and their relationship with process conditions.

[0103] Step 103: Determine the spatial distribution characteristics of salt crystal particles in the saturated solution based on the set of motion trajectories.

[0104] Optionally, step 103 may specifically include the following steps:

[0105] Step 1031: Based on the set of motion trajectories, extract the three-dimensional spatial position sequence of salt crystal particles in the crystallization tank;

[0106] Step 1032: Calculate the distribution density of salt crystallization particles in the axial and radial directions of the crystallization tank based on the three-dimensional spatial position sequence;

[0107] Step 1033: Based on the three-dimensional spatial position sequence, analyze the movement direction of salt crystal particles in the saturated solution, and classify salt crystal particles with the same movement direction into the same movement group;

[0108] Step 1034: Combining the distribution density and the motion group, divide the crystallization tank into a high-density motion area and a low-density motion area.

[0109] Step 1035: Based on the spatial relationship between the high-density motion region and the low-density motion region, determine the spatial distribution characteristics of salt crystal particles in the saturated solution.

[0110] In the above steps, spatial distribution characteristics refer to a comprehensive and structured description of the overall situation of the salt crystal particle group in the three-dimensional space of the crystallization tank, including its quantity, density, aggregation state, and movement trend.

[0111] The three-dimensional spatial position sequence refers to the use of three-dimensional reconstruction technologies such as binocular vision or structured light to convert and fuse the two-dimensional position coordinates of salt crystal particles obtained from two-dimensional images acquired from different perspectives into a series of continuous position points in the real three-dimensional physical coordinate system of the crystallization tank, thus completely describing the movement path of a salt crystal particle in three-dimensional space.

[0112] Distribution density refers to the frequency or quantity of salt crystal particles appearing per unit volume or unit area in the axial and radial directions within a crystallization tank. It is used to quantify the density of particles in space and is obtained by projecting and statistically calculating a three-dimensional spatial position sequence in a specified direction.

[0113] The direction of motion refers to the direction in which the instantaneous velocity of salt crystal particles points when they move in a saturated solution. It is used to describe the trend of the movement of salt crystal particles and is obtained by calculating the difference vector of continuous position points in a three-dimensional spatial position sequence and normalizing it.

[0114] Motion groups refer to grouping salt crystal particles with similar motion trends into the same set based on the consistency of their motion direction. This is used to identify different flow field structures or motion patterns that may exist in the crystallization tank. It is obtained by clustering analysis of the motion directions of a large number of salt crystal particles.

[0115] High-density movement areas refer to the spatial range within the crystallization tank where the distribution density of salt crystal particles is significantly higher than that of the surrounding area, and where the salt crystal particles have a specific and consistent direction of movement. This indicates the existence of salt crystal particle aggregation or specific flow. It is obtained by combining the distribution density map and the movement group map to perform region growth or threshold segmentation.

[0116] Low-density movement zone refers to the spatial range within the crystallization tank where the distribution density of salt crystal particles is significantly lower than that of the surrounding area, or where the movement direction of salt crystal particles is chaotic and inconsistent. It may correspond to areas with weaker flow or fewer salt crystal particles and is delineated by comparison with high-density movement zone.

[0117] In this embodiment, firstly, the three-dimensional spatial coordinates of the two-dimensional image pixel coordinates on each trajectory in the set of motion trajectories are calculated in the three-dimensional spatial coordinate system of the crystallizing tank by applying the triangulation principle through three-dimensional reconstruction technology, thereby obtaining the three-dimensional spatial position sequence.

[0118] Secondly, using the spatial grid statistical method, the three-dimensional space of the crystallizer is divided into a series of small voxels or grid units along its axial and radial directions. The number of times each grid unit is visited by the salt crystallization particle location point during the observation time is counted, and the distribution density of salt crystallization particles in the axial and radial directions of the crystallizer is calculated.

[0119] Next, through vector analysis and clustering algorithms, the motion direction of the three-dimensional spatial position sequence at each time point is calculated, and the directional clustering algorithm is used to group salt crystal particles with similar motion directions into the same motion group;

[0120] Then, through image processing and region segmentation techniques, the axial and radial distribution density maps are fused to generate a three-dimensional density distribution map. At the same time, the information of each motion group is mapped to the three-dimensional space with different colors or labels. In the three-dimensional density distribution map, high-density motion regions with density values ​​that are consistently higher than a certain threshold and with high consistency in the internal particle motion direction, and low-density motion regions with density values ​​that are consistently lower than a certain threshold and with chaotic internal particle motion direction are identified.

[0121] Finally, through spatial relationship modeling and feature extraction, the spatial relationship and structural information of high-density motion regions and low-density motion regions in the crystallization tank are analyzed and integrated to determine the spatial distribution characteristics of salt crystal particles in the saturated solution.

[0122] Following the previous implementation case, the 3D reconstruction module is first invoked, and the precise calibration parameters of the high-speed industrial camera of brand B are used to convert tens of thousands of two-dimensional pixels on each motion trajectory into a three-dimensional spatial position sequence within the crystallization tank.

[0123] The analysis program then virtually divided the 5-meter-high, 3-meter-diameter crystallization tank into millions of tiny cubic grids, counted the number of times each grid was traversed by the trajectory of salt crystal particles during a 10-second observation period, and calculated the axial and radial distribution densities respectively.

[0124] At the same time, the average motion direction of each salt crystal particle is calculated, and all salt crystal particles are divided into several motion groups such as upward motion, slow sinking and horizontal vortex by clustering algorithm;

[0125] Then, the distribution density and motion group are superimposed, and the high-density motion area in the lower part of the crystallization tank with higher density and salt crystal particles generally moving upward is automatically delineated in the three-dimensional model, and the low-density motion area in the upper center of the crystallization tank with lower density and salt crystal particles moving in different directions is delineated.

[0126] Finally, a comprehensive analysis of the spatial relationship between the two regions revealed that the spatial distribution characteristics of salt crystal particles during this observation period were as follows: salt crystal particles were highly concentrated in the upward flow region in the lower part of the crystallization tank, forming the main crystallization zone, while there was a sparsely populated region with relatively calm flow at the top center of the crystallization tank.

[0127] This step utilizes technologies such as 3D reconstruction to transform particle trajectory data into a deep understanding of the macroscopic distribution and movement of particles within the tank. The obtained spatial distribution features reveal the 3D structure of the particles, and by correlating density with movement trends, physical characteristic regions can be identified. This achieves a cognitive leap from the movement of individual particles to the coordinated movement of groups, providing crucial spatial structure and physical background for subsequent correction and precise analysis. It is a core component of intelligent working condition sensing.

[0128] Step 104: Combine the real-time temperature parameters and real-time concentration parameters during the industrial salt evaporation and crystallization process to perform multi-physics field coupling correction on the spatial distribution characteristics to obtain particle distribution information.

[0129] Optionally, step 104 may specifically include the following steps:

[0130] Step 1041: Based on the high-density motion region and low-density motion region in the spatial distribution characteristics, establish an initial particle distribution model in the crystallization tank;

[0131] Step 1042: Correct the settling velocity of salt crystal particles in the initial particle distribution model according to the real-time temperature parameters to obtain the corrected settling velocity distribution.

[0132] Step 1043: Correct the aggregation state of salt crystal particles in the initial particle distribution model according to the real-time concentration parameters to obtain the corrected aggregation state distribution.

[0133] Step 1044: By fusing the corrected settlement velocity distribution and the corrected aggregation state distribution, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours.

[0134] Step 1044 may specifically include the following steps:

[0135] Establish a correspondence between the corrected settling velocity distribution and the corrected aggregation state distribution; based on the correspondence, determine the composite influence parameter on the spatial distribution of particles when the corrected settling velocity distribution and the corrected aggregation state distribution work together; based on the composite influence parameter, calculate the movement vector of each point on the boundary between the high-density motion region and the low-density motion region.

[0136] Specifically, the composite influence parameters are calculated as follows: the boundary contours of the high-density motion region and the low-density motion region are discretized to obtain boundary sampling points. Read from the corrected settlement velocity distribution Settlement velocity at the location Read from the corrected cluster state distribution Aggregation intensity at location According to the preset minimum and maximum values ​​respectively , Normalization and ;according to Calculate composite influence parameters ,in , , For preset weights and + + =1, =1,2,3,...N, where N is the total number of boundary sampling points obtained from discrete sampling of the boundary contour.

[0137] After obtaining the composite influence parameters Then, according to Calculate boundary sampling points The movement vector, where To preset the boundary response coefficients, Let be the boundary normal vector pointing from the low-density motion region to the high-density motion region; when When the value increases, it indicates that the combined effect of decreased settling velocity and increased aggregation intensity is enhanced, and the corresponding boundary sampling points along the [missing information - likely a specific direction or path] are [affected / affected]. The distance traveled in the direction increases.

[0138] The spatial positions of each point on the boundary are adjusted according to the movement vector to form a boundary point set; based on the boundary point set, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours.

[0139] Step 1045: Integrate the updated boundary contour into the initial particle distribution model to output particle distribution information through the integrated particle distribution model.

[0140] In the above steps, the real-time temperature parameter refers to the solution temperature value measured and transmitted in real time by a temperature sensor installed in the crystallization tank during the industrial salt evaporation and crystallization process. It is used to characterize the thermodynamic state that affects the solution viscosity and the intensity of Brownian motion of particles, and is obtained directly through the temperature sensor.

[0141] Real-time concentration parameters refer to the solution mass concentration or supersaturation value measured and transmitted in real time by an online concentration meter during the industrial salt evaporation and crystallization process. They are used to characterize the chemical potential intensity of the solution driving crystal growth and aggregation and are directly obtained through an online concentration sensor.

[0142] Particle distribution information refers to an enhanced, high-fidelity data model that integrates the spatial location, local density, movement trend, and distribution state of salt crystal particles after correction by a physicochemical field. It is used to accurately describe the true distribution of salt crystal particles under the current operating conditions and is obtained by coupling and correcting real-time process parameters to spatial distribution characteristics.

[0143] The initial particle distribution model refers to a three-dimensional digital model that uses spatial distribution characteristics as the core framework to initially describe the density and movement zoning of salt crystallization particles in the crystallization tank. It serves as a benchmark for subsequent corrections and is established based on spatial distribution characteristics.

[0144] The settling velocity of salt crystal particles refers to the final descent velocity of salt crystal particles in a saturated solution when they reach equilibrium under the combined effects of gravity, buoyancy, and fluid resistance. It is a temperature-sensitive function used in models to reflect the influence of temperature on particle mobility.

[0145] The corrected settling velocity distribution refers to the numerical distribution of salt crystallization particle settling velocity at various points in the three-dimensional space of the crystallization tank that is more matched to the current temperature field after local adjustment of the salt crystallization particle settling velocity preset in the initial particle distribution model or calculated according to the general formula based on real-time temperature parameters. It is calculated by inputting real-time temperature parameters into the empirical relationship between settling velocity and temperature.

[0146] The aggregation state of salt crystal particles refers to the degree to which salt crystal particles tend to approach each other, adhere, or form loose clusters due to collisions, van der Waals forces, etc. It is a sensitive function of solution concentration and is used to reflect the influence of concentration on the interaction between salt crystal particles in the initial particle distribution model.

[0147] The corrected aggregation state distribution refers to the numerical distribution of salt crystal particle aggregation intensity at each point in the three-dimensional space of the crystallizer that is more consistent with the current concentration field after local adjustment of the salt crystal particle aggregation tendency parameters preset in the initial particle distribution model based on the real-time concentration parameters. It is calculated by inputting the real-time concentration parameters into the empirical relationship of aggregation state concentration.

[0148] Boundary contour refers to the three-dimensional surface used to separate high-density motion regions from low-density motion regions in the initial particle distribution model, defining the spatial range of different feature regions;

[0149] The updated boundary profile refers to the new shape and position of the separating surface between the high-density motion region and the low-density motion region after the combined influence of the fused and corrected sedimentation velocity distribution and the corrected aggregation state distribution. It reflects the reshaping effect of the physical field on the distribution structure of salt crystal particles and is obtained by applying a displacement derived from the composite influence parameters to the original boundary profile.

[0150] The correspondence refers to the numerical mapping or functional relationship between the settling velocity value given by the corrected settling velocity distribution and the aggregation state value given by the corrected aggregation state distribution at the same point in the crystallization tank. It is used to quantify the joint effect mode of two physical fields at this point and is established through statistical analysis or a model based on physical laws.

[0151] The composite influence parameter refers to a scalar value that comprehensively quantifies the net effect of the combined effect of the corrected settling velocity and the corrected aggregation state at a specific spatial point, causing salt crystal particles to tend to aggregate towards high-density areas or disperse from low-density areas. It is used to drive the deformation of the boundary profile and is calculated by applying a function defined by the correspondence to two corrected distribution values ​​at the same point.

[0152] The movement vector is a three-dimensional spatial vector at each discrete point on the boundary profile, determined by the composite influence parameters at that point. This vector indicates the direction and distance that the boundary point should move in response to changes in the physical field. It is obtained by performing spatial gradient calculations on the composite influence parameters and combining them with the boundary normal.

[0153] The boundary point set refers to the set of all three-dimensional spatial coordinate points that constitute the updated boundary contour. It is a discretized digital representation of the boundary contour, obtained by applying a translation vector to the coordinates of each point on the original boundary contour.

[0154] The integrated particle distribution model refers to the process of reintegrating the new regional structure information contained in the updated boundary contours into the initial particle distribution model, thereby generating a complete and unified particle distribution data model that contains the latest regional divisions and may have updated internal attributes. This is the final particle distribution information, obtained through data replacement and interpolation integration operations.

[0155] In this embodiment, firstly, using three-dimensional modeling technology, the geometric range of high-density and low-density motion regions in the spatial distribution characteristics, as well as the average motion trend of salt crystallization particles within the region, are used as core attributes to construct an initial particle distribution model in the crystallization tank in the computer.

[0156] Secondly, by consulting the physical property database or calling the pre-stored empirical formula, a function relating the settling velocity of salt crystallization particles to temperature is established. The real-time temperature parameters are substituted into this function to calculate the benchmark settling velocity that reflects the average temperature in the current crystallizer. Then, combined with the assumption of possible temperature differences in different regions in the initial particle distribution model, this benchmark value is adjusted to generate the corrected settling velocity distribution.

[0157] Next, the function describing the relationship between particle aggregation tendency and solution concentration is called, and the real-time concentration parameter is substituted into it to calculate the baseline aggregation state parameter. The baseline aggregation state parameter is then differentially allocated within the spatial framework of the initial particle distribution model to generate the corrected aggregation state distribution.

[0158] Next, a correspondence was established between the corrected settling velocity distribution and the corrected aggregation state distribution. Dense sampling was performed on the boundaries of the high-density and low-density regions defined by the initial particle distribution model. Based on the spatial coordinates of each sampling point, the corresponding settling velocity and aggregation state values ​​were read from the corrected settling velocity and aggregation state distributions. The composite influence parameters at the sampling points were calculated based on the correspondence function. Based on the composite influence parameters and the geometric normal direction of the boundary at the sampling points, the movement vector that drives the boundary points to move was calculated. All sampling points on the boundary were traversed, and the spatial coordinates of each point were added to the movement vector to obtain the new positions after movement. These new positions constituted the boundary point set. A surface reconstruction algorithm was used to connect the boundary point set and smooth it into a continuous three-dimensional surface as the updated boundary contour.

[0159] Finally, through data integration and model update operations, the updated boundary contours are re-imported into the initial particle distribution model. Based on the new boundary contours, the attributes of each partition within the model are recalculated and adjusted. The particle distribution information is then output through the integrated particle distribution model.

[0160] Following the implementation case of the previous step, firstly, based on the spatial distribution characteristics, an initial particle distribution model containing high-density motion regions and low-density motion regions is established in the digital space;

[0161] Secondly, based on the real-time temperature parameter of 85°C, it was calculated that the theoretical settling velocity of salt crystal particles under this real-time temperature parameter is slower than the standard value, and a corrected settling velocity distribution with an overall slower settling velocity was generated.

[0162] Next, based on the real-time concentration parameter of 28%, it was determined that the solution was in a high supersaturation state. After the salt crystal particles collided, they were prone to sticking together, resulting in a modified aggregation state distribution that enhanced the tendency of particle aggregation.

[0163] Next, the corrected settling velocity distribution and the corrected aggregation state distribution were integrated. The boundary points of the high-density motion region and the low-density motion region divided by the initial model were checked one by one. It was found that many points in the lower part of the boundary met the conditions of slow settling velocity and strong aggregation state after correction. The calculated composite influence parameters showed that the salt crystal particles at these boundary points had a strong tendency to aggregate into the interior of the high-density region. Therefore, the movement vectors of these boundary points were calculated. After moving all boundary points according to these movement vectors, the boundary contour was reconstructed with the new point set.

[0164] Finally, the new boundary profile was integrated into the initial particle distribution model, and the statistical properties within the new region were recalculated. The final output was a particle distribution information showing that at 85°C and 28% concentration, the actual effective area of ​​high particle aggregation was larger than that observed visually, and particle retention was more significant.

[0165] This step transforms real-time temperature and concentration parameters into quantitative corrections to particle dynamics, driving the evolution of the visual observation model to compensate for the insensitivity of pure visual analysis to changes in implicit physicochemical fields. The generated particle distribution information not only presents particle position and motion but also profoundly reflects the impact of operating conditions, forming a high-fidelity digital mapping that integrates multi-source information. This lays a reliable data foundation for accurate particle size analysis and represents a key intelligent upgrade from observation to perception and understanding.

[0166] Step 105: Generate an initial particle size distribution curve for industrial salt crystallization based on the particle distribution information, and adjust the initial particle size distribution curve to obtain the target particle size distribution curve.

[0167] Optionally, step 105 may specifically include the following steps:

[0168] Step 1051: Based on the spatial location and particle size characteristics of the salt crystal particles in the particle distribution information, count the number of salt crystal particles in different particle size ranges and generate an initial particle size distribution curve.

[0169] Step 1052: Extract the particle size characteristics of salt crystal particles in the high-density motion region and the low-density motion region from the particle distribution information;

[0170] Step 1053: Calculate the adjustment parameters for different particle size segments in the initial particle size distribution curve based on the particle size characteristic differences.

[0171] Step 1054: Apply the adjustment parameters to the corresponding particle size segment of the initial particle size distribution curve to form an intermediate particle size distribution curve;

[0172] Step 1055: Based on the spatial distribution uniformity of salt crystal particles in the particle distribution information, the intermediate particle size distribution curve is smoothed to generate the target particle size distribution curve.

[0173] In the above steps, the initial particle size distribution curve refers to a statistical chart that shows the proportion of the number or mass of salt crystal particles of different sizes in the industrial salt crystallization product at the observation time. The horizontal axis represents the particle size, and the vertical axis represents the number or frequency of particles of that size. It is used to initially reflect the general situation of the particle size composition of the product. It is directly generated by statistically analyzing the particle size characteristics of all salt crystal particles in the particle distribution information.

[0174] The target particle size distribution curve refers to the particle size distribution curve obtained after further correction and optimization based on the initial particle size distribution curve, which incorporates information such as the distribution differences of salt crystal particles in different areas of the crystallization tank and the uniformity of the overall spatial distribution. It is used to more realistically and accurately represent the particle size distribution of industrial salt crystallization products under the current working conditions. It is obtained by adjusting and smoothing the initial particle size distribution curve.

[0175] Spatial location refers to the specific coordinates of salt crystallization particles in the three-dimensional space of the crystallization tank. It is used to determine which feature region each particle is located in and is obtained directly from the geometric model part of the particle distribution information.

[0176] Particle size characteristics refer to attributes used to describe the size of individual salt crystal particles. They are usually expressed as equivalent diameters and are calculated by image analysis techniques based on the projected area or volume of the salt crystal particles. These characteristics are then stored as attributes in the particle distribution information.

[0177] The difference in particle size characteristics refers to the systematic differences in statistical characteristics such as average particle size, main particle size range, or particle size distribution shape between salt crystal particle groups located in high-density motion regions defined by particle distribution information and salt crystal particle groups located in low-density motion regions. It is obtained by statistically analyzing the particle size characteristics of particles in the two regions separately and comparing them.

[0178] Adjustment parameters refer to a set of values ​​calculated based on differences in particle size characteristics. They are used to quantitatively describe the magnitude and direction of increasing or decreasing the values ​​of each particle size segment in the initial particle size distribution curve. This is achieved by quantifying the differences in particle size characteristics into the impact values ​​on each segment of the curve.

[0179] The intermediate granularity distribution curve refers to a transitional granularity distribution curve that has been initially corrected but may still contain local fluctuations or unevenness after the adjustment parameters are applied to the corresponding granularity segment of the initial granularity distribution curve. It is an intermediate result for generating the final target curve.

[0180] Spatial distribution uniformity refers to the degree of homogeneity of salt crystal particles in the three-dimensional space of the crystallization tank. It is used to measure whether the density and particle size composition of the particles are similar in all parts of the space, or whether there is obvious local enrichment or sparseness. It is obtained by calculating the uniformity index of the spatial coordinates of the particle points in the particle distribution information.

[0181] In this embodiment, firstly, the particle size characteristic value contained in each salt crystal particle record is read from the particle distribution information data structure through data query and statistical technology. A series of continuous particle size intervals covering the entire possible particle size range are predefined. Each salt crystal particle is assigned to the corresponding particle size interval according to its particle size characteristic value. The initial particle size distribution curve is generated with the particle size interval as the horizontal axis and the number of salt crystal particles in each interval as the vertical axis.

[0182] Secondly, by using regional filtering and grouping statistics techniques, the spatial location information of each salt crystallization particle recorded in the particle distribution information is utilized. Combined with the geometric boundaries of the high-density movement region and low-density movement region defined in the particle distribution information, the differences in particle size characteristics are calculated by comparing the numerical values, peak positions, and distribution widths in the same particle size range.

[0183] Next, using influence factor mapping and parameter calculation techniques, an adjustment parameter is calculated for each granularity segment of the initial granularity distribution curve;

[0184] Then, the adjustment parameters were found through curve calculation techniques, and the ordinate values ​​of all data points were modified through arithmetic operations. The intermediate granularity distribution curve was then reconnected with the new data points.

[0185] Finally, using smoothing filtering techniques in signal processing, the selected smoothing filter is applied to the ordinate sequence of the intermediate granularity distribution curve, and new, smoother data points are connected to generate the target granularity distribution curve.

[0186] Following the previous implementation case, the equivalent diameter data of tens of thousands of salt crystal particles were first read from the particle distribution information. The number of particles was counted in groups of 5 micrometers, and an initial particle size distribution curve showing the change of particle number with particle size was plotted.

[0187] Secondly, based on the updated regional division in the particle distribution information, the particle size distribution of salt crystal particles located in the expanded high-density motion region and the remaining low-density motion region were statistically analyzed, and the differences in particle size characteristics were obtained by comparison.

[0188] Next, adjustment parameters are calculated based on the differences in particle size characteristics. For particle size ranges smaller than 100 micrometers, the adjustment parameters are negative; for particle size ranges larger than 200 micrometers, the adjustment parameters are small positive values.

[0189] Then, these adjustment parameters are applied to the initial particle size distribution curve to obtain the intermediate particle size distribution curve;

[0190] Finally, the spatial distribution uniformity index of particle points in the current particle distribution information is calculated. It is found that the uniformity is generally poor. Therefore, a medium-strength smoothing filter is selected to smooth the intermediate curve and generate the target particle size distribution curve.

[0191] This step extracts particle size data from particle distribution information to generate an initial curve, and then uses the region division and spatial uniformity information to optimize the curve to obtain the target particle size distribution curve. This method overcomes the limitations of simple overall counting, fully considers the spatial differences and statistical stability of particle distribution within the crystallizer, and makes the results more robust and realistically reflect the actual particle size situation under dynamic crystallization environment, providing a reliable basis for subsequent accurate quality matching.

[0192] Step 106: Perform a matching degree analysis between the target particle size distribution curve and the preset industrial salt crystallization quality standard, and output the online detection results of the industrial salt crystallization particle size distribution.

[0193] Optionally, step 106 may specifically include the following steps:

[0194] Step 1061: Decompose the target particle size distribution curve into multiple target segments according to the particle size interval, and at the same time decompose the preset numerical range of different particle size intervals in the preset industrial salt crystallization quality standard into multiple standard segments.

[0195] Step 1062: Based on the target segment and the standard segment, compare the distribution value difference between each target segment and the corresponding standard segment one by one to obtain the matching deviation of each granularity segment;

[0196] Step 1063: Based on the matching deviation of each particle size segment, a weighted comprehensive calculation is performed in combination with the weight values ​​preset for different particle size segments in the industrial salt crystallization quality standard to obtain the comprehensive matching degree.

[0197] Step 1064: Compare the overall matching degree with the preset qualified threshold, and determine the quality grade of the industrial salt crystallization product based on the comparison result;

[0198] Step 1065: Combine the key feature parameters extracted from the target particle size distribution curve with the quality level to generate online detection results.

[0199] In the above steps, the preset industrial salt crystallization quality standard refers to the quantitative requirements that the particle size distribution of industrial salt crystallization products should meet as predefined by the manufacturer or product specifications. It usually specifies the allowable numerical range of particle quantity or mass percentage in different particle size ranges, and is used as a benchmark for judging whether the product is qualified. It is pre-stored and called through a database or configuration file.

[0200] The online detection result refers to the comprehensive report generated by this method after performing online particle size analysis on the current batch of industrial salt crystallization products. It includes quality judgment conclusions, detailed distribution data and key characteristic descriptions. It is used to guide production operations or quality judgment in real time. The report is generated by comparing and integrating the particle size data obtained from the analysis with preset standards.

[0201] The target segment refers to a series of continuous, non-overlapping granularity range segments into which the target granularity distribution curve is cut along its horizontal axis according to a pre-defined granularity interval. Each segment represents a specific granularity interval and is obtained by segmenting the target granularity distribution curve.

[0202] The preset numerical range refers to the minimum and maximum values ​​that a qualified product should achieve in terms of the number of particles or the mass percentage within each specific particle size range in the preset industrial salt crystallization quality standard. It defines the qualified window for each particle size range and reads it from the data structure of the quality standard.

[0203] The standard segment refers to a particle size interval data unit that corresponds one-to-one with the target segment, extracted from the preset industrial salt crystallization quality standard, and contains a preset numerical range. It is used for direct comparison with the target segment and is obtained by decomposing the numerical range in the quality standard according to the same particle size interval division rules.

[0204] Distribution numerical difference refers to the degree of deviation between the actual measured value represented by the target segment and the preset numerical range specified by the standard segment corresponding to the same granularity interval. It is usually quantified as the distance of the actual value from the standard range or the difference from the standard median value, and is obtained by comparing the two values ​​arithmetically.

[0205] The matching deviation of each granularity segment refers to the result of standardization or formatting of the difference in distribution values ​​calculated for each granularity interval. It is a quantitative single value that represents the degree of mismatch between the actual product distribution and the standard requirements in that specific granularity interval. It is obtained by calculating the difference in distribution values ​​and possibly performing subsequent processing.

[0206] The preset weight value refers to the importance coefficient assigned to different particle size ranges in the preset industrial salt crystallization quality standard. It reflects the magnitude of the influence of particle distribution within the particle size range on the final product quality. The larger the weight value, the more important the range. It is read from the data structure of the quality standard.

[0207] The overall matching degree refers to a single, comprehensive numerical indicator. It is obtained by weighting and summing the matching deviations of all granularity segments according to their corresponding preset weight values. It is used to quantify the degree of conformity between the granularity distribution of the current product and the quality standard. The higher the value, the better the conformity. It is obtained through weighted calculation.

[0208] The preset pass threshold is a pre-set critical value used to determine whether the overall matching degree meets the pass requirements. If the calculated overall matching degree reaches or exceeds this threshold, the overall particle size distribution of the product is considered to be qualified; otherwise, it is unqualified. This threshold is preset according to the product grade requirements.

[0209] The comparison result refers to the logical judgment conclusion obtained by comparing the calculated comprehensive matching degree with the preset qualified threshold. It may also include a specific description of the degree to which the threshold is exceeded, which is obtained through simple logical comparison operations.

[0210] The quality grade refers to the final qualitative evaluation grade assigned to the current industrial salt crystallization product regarding its particle size distribution quality based on the comparison results and in combination with possible grading rules. It is obtained by mapping the comparison results to preset grading rules.

[0211] Key feature parameters refer to several key values ​​extracted from the target granularity distribution curve that can summarize the core shape and statistical characteristics of the curve. They are used to describe the specific characteristics of the product granularity distribution and are obtained by feature extraction calculation of the target granularity distribution curve.

[0212] In this embodiment, firstly, the particle size distribution curve is uniformly divided into a series of continuous target segments by using data segmentation technology according to a pre-set, fixed particle size interval width. Simultaneously, the allowable distribution value range of each particle size interval is read from the preset industrial salt crystallization quality standard according to the same particle size interval definition as the target segment division, thereby forming a series of multiple standard segments corresponding to the target segments.

[0213] Secondly, by comparing and quantifying the differences element by element, the actual measured value represented by each target segment is compared with the value range specified by the corresponding standard segment. The deviation of the actual measured value from the standard range in each granularity interval is calculated, thereby obtaining the matching deviation of each granularity segment.

[0214] Next, using a weighted comprehensive evaluation model, based on the importance weight values ​​pre-set for different particle size ranges in the industrial salt crystallization quality standard, the matching deviation of all particle size ranges is weighted and calculated to obtain the comprehensive matching degree value.

[0215] Then, through threshold comparison logic, the calculated comprehensive matching degree is compared with the preset threshold, and the quality grade corresponding to the current industrial salt crystallization product is determined based on the comparison result.

[0216] Finally, using report generation and information fusion technology, key feature parameters such as main distribution granularity and distribution width are extracted from the target granularity distribution curve, and these key feature parameters are combined with the quality level to generate online detection results.

[0217] Following the implementation case of the previous step, firstly, the target particle size distribution curve is divided into multiple target segments at intervals of 10 micrometers, and the allowable content range of the corresponding particle size interval is read from the preset superior industrial salt quality standard file at the same time, and the preset value range of different particle size intervals is decomposed into multiple standard segments.

[0218] Secondly, the comparison is performed segment by segment. The comparison standard is that in the 150-160 micrometer range, the target curve shows a content of 15%, while the standard range is 12%-18%, so the matching deviation in this range is 0; in the 80-90 micrometer range, the target content is 8%, but the standard upper limit is 6%, so a positive deviation is calculated. The matching deviation is calculated for all particle size segments.

[0219] Next, the weighted average of the matching deviations of all particle size ranges is calculated by calling the preset weight values ​​in the preset industrial salt crystallization quality standard, and the overall matching degree is 0.92.

[0220] The overall matching degree is then compared with the preset qualified threshold. The preset qualified threshold is 0.90 for superior products and 0.80 for qualified products. Since 0.92 is greater than 0.90, the quality grade of this batch of industrial salt crystal products is determined to be superior.

[0221] Finally, key feature parameters are extracted from the target granularity distribution curve, and these key feature parameters are combined with the quality level and overall matching degree to generate and output online detection results.

[0222] This step involves precise matching and weighted evaluation of the target particle size distribution curve against preset standards, enabling automated online determination of particle size quality. This method overcomes the limitations of manual judgment, comprehensively measuring compliance through integrated matching and determining the quality level based on clear thresholds. The final output test results contain clear conclusions and key distribution parameters, providing immediate and reliable decision-making support for quality control and process adjustment, completing the closed loop from measurement and analysis to quality evaluation.

[0223] Figure 2 This application provides a schematic diagram of the structure of an online detection system for the particle size distribution of industrial salt crystals based on machine vision, as shown below. Figure 2 As shown, the system includes:

[0224] The acquisition module 21 is used to acquire dynamic image sequences during the evaporation and crystallization process of industrial salt;

[0225] Tracking module 22 is used to track the trajectory of salt crystal particles in the dynamic image sequence to obtain a set of motion trajectories;

[0226] The determination module 23 is used to determine the spatial distribution characteristics of salt crystal particles in a saturated solution based on the set of motion trajectories.

[0227] Correction module 24 is used to combine real-time temperature parameters and real-time concentration parameters during the industrial salt evaporation and crystallization process to perform multi-physics field coupling correction on the spatial distribution characteristics to obtain particle distribution information;

[0228] The adjustment module 25 is used to generate an initial particle size distribution curve for industrial salt crystallization based on the particle distribution information, and to adjust the initial particle size distribution curve to obtain a target particle size distribution curve.

[0229] Output module 26 is used to perform a matching degree analysis between the target particle size distribution curve and the preset industrial salt crystallization quality standard, and output the online detection results of the industrial salt crystallization particle size distribution.

[0230] Figure 2 The aforementioned machine vision-based online detection system for industrial salt crystal particle size distribution can perform... Figure 1 The implementation principle and technical effects of the online detection method for industrial salt crystal particle size distribution based on machine vision described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the online detection system for industrial salt crystal particle size distribution based on machine vision in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0231] In one possible design, Figure 2 The machine vision-based online detection system for industrial salt crystal particle size distribution shown in the embodiment can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0232] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0233] The processing component 32 is used for the above Figure 1 The embodiment describes an online detection method for the particle size distribution of industrial salt crystals based on machine vision.

[0234] The processing component 32 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.

[0235] 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.

[0236] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0237] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0238] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0239] 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.

[0240] 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 an online detection method for the particle size distribution of industrial salt crystals based on machine vision.

[0241] 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.

[0242] 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.

[0243] 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.

[0244] 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 machine vision-based online detection method for the particle size distribution of industrial salt crystals, characterized in that, include: Acquire dynamic image sequences during the evaporation and crystallization process of industrial salt; The motion trajectory set is obtained by tracing the trajectory of salt crystal particles in the dynamic image sequence. Based on the set of motion trajectories, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined; By combining real-time temperature and concentration parameters during the industrial salt evaporation and crystallization process, the spatial distribution characteristics are corrected using multi-physics field coupling to obtain particle distribution information. An initial particle size distribution curve for industrial salt crystallization is generated based on the particle distribution information, and the initial particle size distribution curve is adjusted to obtain the target particle size distribution curve. The target particle size distribution curve is matched with the preset industrial salt crystallization quality standard, and the online detection results of the industrial salt crystallization particle size distribution are output.

2. The method according to claim 1, characterized in that, Trajectory tracking of salt crystal particles is performed on the dynamic image sequence to obtain a set of motion trajectories, including: Identify the position coordinates of salt crystal particles within each image of the dynamic image sequence; The position coordinates of salt crystal particles in two adjacent images in the dynamic image sequence are correlated to determine the positional changes of the salt crystal particles; The positional changes of the salt crystal particles on multiple consecutive dynamic images are connected to form an initial motion trajectory; The initial motion trajectory was identified, and the target trajectory was found to be abnormal. The abnormality of the target trajectory is corrected, and the corrected initial motion trajectory is integrated with the initial motion trajectory that is not affected by the abnormality of the target trajectory to form a motion trajectory set.

3. The method according to claim 1, characterized in that, Based on the set of motion trajectories, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined, including: Based on the set of motion trajectories, the three-dimensional spatial position sequence of salt crystal particles in the crystallization tank is extracted; Based on the three-dimensional spatial position sequence, calculate the distribution density of salt crystal particles in the axial and radial directions of the crystallization tank; Based on the three-dimensional spatial position sequence, the movement direction of salt crystal particles in the saturated solution is analyzed, and salt crystal particles with the same movement direction are divided into the same movement group. Based on the distribution density and the motion group, a high-density motion zone and a low-density motion zone are divided within the crystallization tank; Based on the spatial relationship between the high-density motion region and the low-density motion region, the spatial distribution characteristics of salt crystal particles in a saturated solution are determined.

4. The method according to claim 1, characterized in that, By combining real-time temperature and concentration parameters during the industrial salt evaporation and crystallization process, multi-physics coupling correction is applied to the spatial distribution characteristics to obtain particle distribution information, including: Based on the high-density and low-density motion regions in the spatial distribution characteristics, an initial particle distribution model is established within the crystallization tank. The settling velocity of salt crystal particles in the initial particle distribution model is corrected based on real-time temperature parameters to obtain the corrected settling velocity distribution. The aggregation state of salt crystal particles in the initial particle distribution model is corrected based on the real-time concentration parameters to obtain the corrected aggregation state distribution. By fusing the corrected sedimentation velocity distribution and the corrected aggregation state distribution, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours. The updated boundary contours are integrated into the initial particle distribution model to output particle distribution information through the integrated particle distribution model.

5. The method according to claim 4, characterized in that, By fusing the corrected settlement velocity distribution and the corrected aggregation state distribution, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours, including: Establish the correspondence between the corrected settlement velocity distribution and the corrected aggregation state distribution; Based on the aforementioned correspondence, the combined influence parameters on the spatial distribution of particles when the modified settling velocity distribution and the modified aggregation state distribution work together are determined. Based on the composite influence parameters, calculate the movement vector of each point on the boundary between the high-density motion region and the low-density motion region; Adjust the spatial position of each point on the boundary according to the movement vector to form a boundary point set; Based on the set of boundary points, the boundary contours of the high-density motion region and the low-density motion region are updated to obtain the updated boundary contours.

6. The method according to claim 1, characterized in that, Based on the particle distribution information, an initial particle size distribution curve for industrial salt crystallization is generated, and the initial particle size distribution curve is adjusted to obtain the target particle size distribution curve, including: Based on the spatial location and particle size characteristics of the salt crystal particles in the particle distribution information, the number of salt crystal particles in different particle size ranges is counted to generate an initial particle size distribution curve. Extract the differences in particle size characteristics between salt crystal particles in high-density motion regions and low-density motion regions from the particle distribution information; Based on the differences in particle size characteristics, calculate the adjustment parameters for different particle size segments in the initial particle size distribution curve; The adjustment parameters are applied to the corresponding particle size segment of the initial particle size distribution curve to form an intermediate particle size distribution curve; Based on the spatial uniformity of salt crystal particles in the particle distribution information, the intermediate particle size distribution curve is smoothed to generate the target particle size distribution curve.

7. The method according to claim 1, characterized in that, The target particle size distribution curve is compared with a preset industrial salt crystallization quality standard to analyze its matching degree, and the online detection results of the industrial salt crystallization particle size distribution are output, including: The target particle size distribution curve is decomposed into multiple target segments according to the particle size interval, and the preset numerical range of different particle size intervals in the preset industrial salt crystallization quality standard is decomposed into multiple standard segments. Based on the target segment and the standard segment, the distribution value difference between each target segment and the corresponding standard segment is compared one by one to obtain the matching deviation of each granularity segment; Based on the matching deviation of each particle size segment, a weighted comprehensive calculation is performed using the weight values ​​preset for different particle size segments in the industrial salt crystallization quality standard to obtain the comprehensive matching degree. The overall matching degree is compared with a preset qualified threshold, and the quality grade of the industrial salt crystallization product is determined based on the comparison result. The key feature parameters extracted from the target particle size distribution curve are combined with the quality level to generate online detection results.

8. An online detection system for the particle size distribution of industrial salt crystals based on machine vision, characterized in that, include: The acquisition module is used to acquire dynamic image sequences during the industrial salt evaporation and crystallization process; The tracking module is used to track the trajectory of salt crystal particles in the dynamic image sequence to obtain a set of motion trajectories; The determination module is used to determine the spatial distribution characteristics of salt crystal particles in a saturated solution based on the set of motion trajectories. The correction module is used to combine real-time temperature parameters and real-time concentration parameters during the industrial salt evaporation and crystallization process to perform multi-physics field coupling correction on the spatial distribution characteristics to obtain particle distribution information. The adjustment module is used to generate an initial particle size distribution curve for industrial salt crystallization based on the particle distribution information, and to adjust the initial particle size distribution curve to obtain the target particle size distribution curve. The output module is used to perform a matching degree analysis between the target particle size distribution curve and the preset industrial salt crystallization quality standard, and output the online detection results of the industrial salt crystallization particle size distribution.

9. 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 the online detection method for industrial salt crystallization particle size distribution based on machine vision as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements an online detection method for industrial salt crystal particle size distribution based on machine vision as described in any one of claims 1 to 7.

Images