Intelligent drying and polishing system for pelletized seeds

Abstract

The invention relates to the technical field of pellet drying and polishing, in particular to an intelligent drying and polishing system for pelletized seeds. The system comprises a rotating module, a first image acquisition module, a drying stage identification module, a second image acquisition module and a parameter optimization module. According to the method, the surface images of the pellets are captured in the drying stage, real-time and nondestructive monitoring of the drying process is achieved, the high-moisture stage, the critical stage and the slow drying stage are automatically recognized by analyzing the peak value and the descent rate of a characteristic curve, and hot air temperature switching is accurately controlled; meanwhile, by judging the stability of the drying process in real time, when the polishing breakage rate exceeds the standard, the standard descent rate difference threshold value of the next batch of drying is automatically increased, so that the drying process enters a mild slow drying mode earlier, and cross-batch process parameter self-learning and continuous optimization are achieved; the consistency and the yield are improved, microcracks and damage caused by drying over-excitation are prevented from the source, and the physical integrity of the pellets is improved.

Description

Technical Field

[0001] This invention relates to the field of pellet drying and polishing technology, and more particularly to an intelligent drying and polishing system for pelleted seeds. Background Technology

[0002] In the field of pelleted seed processing, drying and polishing are key processes that determine the final physical quality and sowing suitability of the pellets. Traditional drying processes mainly rely on fixed time and temperature curves and the operator's experience to judge the degree of moisture evaporation, lacking real-time and accurate online monitoring methods for the moisture state of the pellet surface, which easily leads to insufficient or excessive drying. Insufficient drying will cause the pellets to stick and clump together more during subsequent polishing due to surface stickiness; while excessive drying will easily cause the pellet surface to harden and produce microcracks. These hidden defects are exposed or even aggravated during the polishing stage due to mechanical friction, resulting in increased surface damage rate and dust, which seriously affects the strength, flowability and coating effectiveness of the pellets. Existing polishing processes are also mostly based on fixed parameters and cannot be adaptively adjusted according to the actual quality output of the upstream drying process, with the two major links of drying and polishing being isolated from each other. Furthermore, traditional methods struggle to detect and intervene in abnormal situations such as pellet adhesion and breakage during production, and cannot achieve cross-batch process parameter self-optimization based on final product quality feedback. Therefore, there is an urgent need for an intelligent system that can sense the process status in real time, make intelligent decisions on control parameters, and achieve closed-loop linkage between drying and polishing quality. This system would break through the reliance of existing technologies on human experience and improve the stability, automation level, and finished product quality of pellet seed processing.

[0003] Chinese Patent Publication No. CN107318307A discloses a pellet drying and polishing device, comprising: a drying cylinder and a hot air device connected to the bottom of the drying cylinder, wherein at least one air-drying net is provided in the middle of the drying cylinder, the circumferential edge of the air-drying net is in contact with the inner wall of the drying cylinder, the lower part of the drying cylinder is a flow equalization chamber, wherein at least one flow equalization plate is provided in the flow equalization chamber, and the flow equalization plate has a plurality of flow diversion holes; it can be seen that the existing intelligent drying and polishing technology for pelleted seeds lacks data related to the drying stage, resulting in high micro-cracks and breakage rates in the pellets. Summary of the Invention

[0004] Therefore, the present invention provides an intelligent drying and polishing system for pelleted seeds to overcome the problem in the prior art where the drying and polishing processes are disconnected, resulting in high micro-cracks and breakage rates in the pellets, leading to low physical integrity and processing quality stability.

[0005] To achieve the above objectives, the present invention provides an intelligent drying and polishing system for pelleted seeds, comprising: The rotating module is used for drying and polishing the pellets to be processed; The first image acquisition module, which is connected to the rotation module, is used to monitor the surface morphology image of the pellets to be processed during the drying stage and determine the moisture characteristic curve of the pellets to be processed based on the surface morphology image. The drying stage identification module, which is connected to the first image acquisition module, is used to identify the drying stop point based on the moisture characteristic curve, or to output a fault warning prompt. The second image acquisition module, which is connected to the rotation module, is used to monitor the surface flow image of the pellets to be processed during the polishing stage. The parameter optimization module, which is connected to the drying stage identification module and the second image acquisition module respectively, is used to determine whether to adjust the standard descent rate of the next batch of drying process based on the analysis results of the surface flow image.

[0006] Furthermore, the first image acquisition module includes a shooting unit and a curve drawing unit; The imaging unit is used to periodically capture images of the surface morphology of the pellets to be processed. The curve plotting unit is used to obtain surface gloss and color saturation based on the surface morphology image, calculate image feature values ​​based on surface gloss and color saturation, plot the curve of image feature value change over time, and obtain the moisture characteristic curve.

[0007] Furthermore, the drying stage identification module includes a drying progress analysis unit, a fault identification unit, and a category identification unit; The drying progress analysis unit is used to analyze the moisture characteristic curve and obtain the first analysis result, the second analysis result, and the third analysis result. The fault identification unit is used to respond to the second analysis result, obtain the real-time maximum rate difference of the moisture characteristic curve, analyze the fluctuation of the drying process based on the comparison result of the standard rate difference and the real-time maximum rate difference, and then determine whether to issue a fault warning. The category identification unit is used to identify the cause of the fault and provide a fault warning when the real-time maximum descent rate difference is greater than the standard descent rate difference.

[0008] Furthermore, the drying stage identification module also includes a node determination unit; The node determination unit is used to respond to the real-time maximum descent rate difference being less than or equal to the standard descent rate difference without issuing a warning, until the absolute value of the real-time descent rate is less than or equal to the absolute value of the minimum descent rate, and then determine the current moment as the drying stop node.

[0009] Furthermore, the first analysis result shows that the moisture characteristic curve has no peak, indicating that the current drying stage is in a high moisture stage; The second analysis result shows that the moisture characteristic curve has a peak, and the absolute value of the real-time decline rate is greater than the absolute value of the standard decline rate, indicating that it has entered the critical drying stage. The third analysis result shows that the moisture characteristic curve has a peak, and the absolute value of the real-time descent rate is less than or equal to the absolute value of the standard descent rate, so the hot air temperature is switched to slow drying mode.

[0010] Furthermore, the fault identification unit includes a descent rate calculation subunit; The descent rate calculation subunit is used to calculate several real-time descent rate differences and the real-time maximum descent rate difference based on the moisture characteristic curve; Specifically, for the current sampling period, the image feature values ​​of each sampling point on the moisture characteristic curve are obtained, the rate of decrease of the image feature value corresponding to each sampling point per unit time is calculated to obtain the corresponding real-time decrease rate, the difference between the real-time decrease rates of two adjacent sampling points is calculated to obtain the corresponding real-time decrease rate difference, and the real-time decrease rate with the largest absolute value among the real-time decrease rate differences is selected as the real-time maximum decrease rate difference.

[0011] Furthermore, the fault identification unit also includes a fluctuation analysis subunit; The fluctuation analysis subunit is used to compare the standard descent rate difference with the real-time maximum descent rate difference: If the difference between the real-time maximum descent rate and the standard descent rate is greater than the difference between the two, it is determined that the drying process is fluctuating drastically and a fault warning is issued. If the difference between the real-time maximum descent rate and the standard descent rate is less than or equal to the difference between the two rates, the drying process is considered stable and no fault warning is issued.

[0012] Furthermore, the category identification unit includes a first category identification subunit and a second category identification subunit; The first category identification subunit is used to extract the contour features of the particle outline based on the surface morphology image, determine whether there is particle adhesion based on the contour features, and output a warning prompt to increase the rotation speed or hot air temperature when it exists. The second category identification subunit is used to identify the damage to the surface of the pellets to be processed, so as to determine whether to output a warning prompt to reduce the rotation speed and hot air temperature.

[0013] Furthermore, the parameter optimization module includes a polishing quality analysis unit and a dynamic adjustment unit; The polishing quality analysis unit is used to analyze the actual percentage of broken particles during the polishing stage and compare the actual percentage of broken particles with the allowable percentage of broken particles. The dynamic adjustment unit is used to increase the absolute value of the standard descent rate for the next batch drying process in response to the actual percentage of broken particles being greater than the allowable percentage of broken particles.

[0014] Furthermore, the rotation module includes: The rotary drum has switchable lifting plates on its inner wall. These plates lift the material during drying and replace them with smooth liners during polishing. A drive unit is used to drive the rotary cylinder to rotate; A hot air supply unit is used to control the temperature, humidity, and air volume of the hot air inside the rotating cylinder.

[0015] Compared with existing technologies, the advantages of this invention lie in the fact that, through the first image acquisition module capturing images of the pellet surface during the drying stage, and constructing a moisture characteristic curve based on the ratio of image gloss to saturation, real-time, non-destructive monitoring of the drying process is achieved. The drying stage identification module automatically identifies high moisture, critical, and slow-drying stages by analyzing the peak value and descent rate of the characteristic curve, and precisely controls the switching of hot air temperature. Simultaneously, by calculating the difference in descent rate and comparing it with a preset threshold, the stability of the drying process is determined in real time. When fluctuations exceed limits, an early warning is issued, and the root causes of adhesion or breakage are analyzed. Furthermore, through the second image acquisition module during the polishing stage… The system analyzes the proportion of broken particles in the flow image and feeds this quality result back to the parameter optimization module. When the polishing breakage rate exceeds the standard, the system automatically increases the threshold of the standard descent rate difference for the next batch of drying, allowing the drying process to enter a gentle slow-drying mode earlier. This enables self-learning and continuous optimization of process parameters across batches, improving consistency and yield. It also prevents micro-cracks and breakage caused by excessive drying from the source, improves the physical integrity and seeding applicability of the pellets, and continuously converges to the optimal process window through continuous data accumulation and parameter tuning. While steadily improving quality, it also effectively reduces energy consumption and seed vigor damage. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the intelligent drying and polishing system for pelleted seeds according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the first image acquisition module according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the drying stage identification module according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the fault identification unit in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0018] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0019] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0020] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Please see Figure 1 The diagram shown is a structural schematic of an intelligent drying and polishing system for pelleted seeds according to an embodiment of the present invention. The present invention provides an intelligent drying and polishing system for pelleted seeds, comprising: The rotating module is used for drying and polishing the pellets to be processed; The first image acquisition module, which is connected to the rotation module, is used to monitor the surface morphology image of the pellets to be processed during the drying stage and determine the moisture characteristic curve of the pellets to be processed based on the surface morphology image. The drying stage identification module, which is connected to the first image acquisition module, is used to identify the drying stop point based on the moisture characteristic curve, or to output a fault warning prompt. The second image acquisition module, which is connected to the rotation module, is used to monitor the surface flow image of the pellets to be processed during the polishing stage. The parameter optimization module, which is connected to the drying stage identification module and the second image acquisition module respectively, is used to determine whether to adjust the standard descent rate of the next batch of drying process based on the analysis results of the surface flow image.

[0022] The first image acquisition module captures images of the pellet surface during the drying stage and constructs a moisture characteristic curve based on the ratio of image gloss to saturation, enabling real-time, non-destructive monitoring of the drying process. The drying stage identification module automatically identifies high moisture, critical, and slow-drying stages by analyzing the peak value and descent rate of the characteristic curve, and precisely controls the hot air temperature switching. Simultaneously, by calculating the difference in descent rates and comparing it with a preset threshold, the module determines the stability of the drying process in real time, issuing warnings when fluctuations exceed limits and triggering analysis of the root causes of adhesion or breakage. The second image acquisition module analyzes the breakage in the flow image during the polishing stage. The particle size distribution is measured, and this quality result is fed back to the parameter optimization module. When the polishing breakage rate exceeds the standard, the system automatically increases the threshold of the standard descent rate difference for the next batch of drying, allowing the drying process to enter a gentle slow-drying mode earlier. This enables self-learning and continuous optimization of process parameters across batches, improving consistency and yield. It also prevents micro-cracks and breakage caused by excessive drying from the source, improves the physical integrity and sowing applicability of the pellets, and continuously converges to the optimal process window through continuous data accumulation and parameter tuning. While steadily improving quality, it also effectively reduces energy consumption and seed vigor damage.

[0023] The rotation module in this embodiment includes: The rotary drum has switchable lifting plates on its inner wall. These plates lift the material during drying and replace them with smooth liners during polishing. A drive unit is used to drive the rotary cylinder to rotate; A hot air supply unit is used to control the temperature, humidity, and air volume of the hot air inside the rotating cylinder.

[0024] See Figure 2 As shown, it is a schematic diagram of the structure of the first image acquisition module in an embodiment of the present invention; Specifically, the first image acquisition module includes a shooting unit and a curve drawing unit; The imaging unit is used to periodically capture images of the surface morphology of the pellets to be processed. The curve plotting unit is used to obtain surface gloss and color saturation based on the surface morphology image, calculate image feature values ​​based on surface gloss and color saturation, plot the curve of image feature value change over time, and obtain the moisture characteristic curve. Among them, the image feature value is the ratio of surface gloss to color saturation.

[0025] In this embodiment, the drying stage not only removes moisture from the surface of the pellets but also stabilizes their structure. By precisely controlling temperature, humidity, and airflow, the moisture in the coating layer evaporates slowly. Precise control of temperature and drying intensity is crucial; slow drying prevents rapid surface crusting, internal steam bursting through the pellets leading to cracking, or high temperatures damaging the seed embryo. The polishing stage helps the pellets become more spherical or near-spherical, improving size uniformity and enabling precision seeders to pick up individual seeds without jamming. Polishing also reduces micropores on the pellet surface, forming a smooth "skin," and reduces hygroscopicity, which is beneficial for storage. Furthermore, reduce the premature shedding of pesticides or seed coatings before sowing. If the standard descent rate threshold in the drying process is set "too aggressively," causing excessive or rapid drying, the pellet surface may harden and develop microcracks. These microcracks and brittle surfaces generated during the drying stage will be exposed or even exacerbated by friction during the polishing stage, increasing the surface breakage rate after polishing. Therefore, it is necessary to detect whether quality defects are caused by problems left over from drying during the polishing stage, and then determine whether to increase the standard descent rate of the drying process to make the drying process more gentle and stable, that is, to allow the system to enter the slow drying mode earlier and avoid violent drying.

[0026] Since moisture content is highly correlated with the surface gloss and color saturation of pellets, the moisture content can be inferred by monitoring the changes in optical characteristics during the drying stage. Pellet surfaces with high moisture content have a liquid water film, producing specular reflection and appearing as locally bright "spots" under uniform light. The overall color appears deeper and more saturated due to the absorption of light by water molecules. In pellets with low moisture content / meeting standards, the water film disappears, light becomes diffuse reflection, the gloss weakens, and the color reverts to the relatively lighter hue of the dried coating material itself. An industrial camera is placed at a fixed position inside the drying cylinder, along with a uniform ring-shaped LED light source to reduce shadow interference. Static samples are photographed under uniform light. In the HSV color space, high-brightness and low-saturation pixel areas are extracted from the image, and their area ratio is calculated as the surface gloss. Theoretically, this ratio gradually decreases as drying progresses. The high-brightness threshold is set to V=200 (78% brightness), and the low-saturation threshold is set to S=40 (6% saturation). For example, pixel block A: V=240>200 and S=20<40. The image meets the criteria of high brightness and low saturation. Pixel block B: V=180<200, which does not meet the criteria. Traversing all pixel blocks in the entire image, the total number of pixels meeting this condition is 3200. The total number of pixels in the image is 40000 (200x200). Therefore, the surface gloss = 3200 / 40000 = 0.08. The average saturation is calculated as the average S-channel value of the entire image. The average saturation gradually decreases from its peak as moisture decreases. For example, the sum of the saturation S-values ​​of all pixels in the entire image is 3400000, and the total number of pixels in the image is 40000. 0000, color saturation = 3400000 / 40000 = 85. This value represents the vividness of the color on the surface of the pellet. Moisture will make the color darker and more vivid (the S value increases), and after drying, the color will lighten back to its original color (the S value decreases to a stable value, such as around 70). When calculating the image feature value, the saturation and other indicators need to be normalized to between 0 and 1, that is, the color saturation is divided by 255. For example, dividing the color saturation of 85 by 255 converts it to 0.33, then the image feature value = 0.08 / 0.33 ≈ 0.24.

[0027] See Figure 3 As shown, it is a structural schematic diagram of the drying stage identification module in an embodiment of the present invention; Specifically, the drying stage identification module includes a drying progress analysis unit, a fault identification unit, and a category identification unit; The drying progress analysis unit is used to analyze the moisture characteristic curve and obtain the first analysis result, the second analysis result, and the third analysis result. The fault identification unit is used to respond to the second analysis result, obtain the real-time maximum rate difference of the moisture characteristic curve, analyze the fluctuation of the drying process based on the comparison result of the standard rate difference and the real-time maximum rate difference, and then determine whether to issue a fault warning. The category identification unit is used to identify the cause of the fault and provide a fault warning when the real-time maximum descent rate difference is greater than the standard descent rate difference.

[0028] Specifically, the drying stage identification module also includes a node determination unit; The node determination unit is used to respond to the real-time maximum descent rate difference being less than or equal to the standard descent rate difference without issuing a warning, until the absolute value of the real-time descent rate is less than or equal to the absolute value of the minimum descent rate, and then determine the current moment as the drying stop node.

[0029] In this embodiment, the minimum descent rate is set to -0.005.

[0030] Specifically, the first analysis result shows that the moisture characteristic curve has no peak, indicating that the current drying stage is in a high moisture stage. The second analysis result shows that the moisture characteristic curve has a peak, and the absolute value of the real-time decline rate is greater than the absolute value of the standard decline rate, indicating that it has entered the critical drying stage. The third analysis result shows that the moisture characteristic curve has a peak, and the absolute value of the real-time descent rate is less than or equal to the absolute value of the standard descent rate, so the hot air temperature is switched to slow drying mode.

[0031] In this embodiment, the specific process of switching the hot air temperature to the slow drying mode is as follows: the hot air temperature is switched from 60℃ to 40℃ for the slow drying mode; the standard descent rate is set to -0.03; when the moisture characteristic curve has a peak, the absolute value of the real-time descent rate is obtained in time to analyze the speed of moisture evaporation. A larger absolute value corresponds to a faster descent. If it is determined that the absolute value of the real-time descent rate is greater than the absolute value of the standard descent rate, it indicates that moisture evaporation is still very fast and there is still a lot of free water that can be quickly removed. In this case, the current drying intensity is maintained and rapid dehydration continues. At the same time, the drying quality is monitored based on the difference between the real-time maximum descent rates to diagnose whether the drying power is uniform and stable during the rapid dehydration process, and whether there are local over-wet or overheated areas that cause adhesion or damage. If it is determined that the absolute value of the real-time descent rate is less than or equal to the absolute value of the standard descent rate, it indicates that the moisture evaporation rate has decreased significantly, and the remaining water is mostly bound water that is difficult to remove. If strong drying is continued, the energy efficiency is low and the seeds are easily damaged. The rapid dehydration mode needs to be switched to the mild slow drying mode to prevent excessive drying and cracking in the later stage.

[0032] See Figure 4 As shown, it is a structural schematic diagram of the fault identification unit in an embodiment of the present invention; Specifically, the fault identification unit includes a descent rate calculation subunit; The descent rate calculation subunit is used to calculate several real-time descent rate differences and the real-time maximum descent rate difference based on the moisture characteristic curve; Specifically, for the current sampling period, the image feature values ​​of each sampling point on the moisture characteristic curve are obtained, the rate of decrease of the image feature value corresponding to each sampling point per unit time is calculated to obtain the corresponding real-time decrease rate, the difference between the real-time decrease rates of two adjacent sampling points is calculated to obtain the corresponding real-time decrease rate difference, and the real-time decrease rate with the largest absolute value among the real-time decrease rate differences is selected as the real-time maximum decrease rate difference.

[0033] Specifically, the fault identification unit further includes a fluctuation analysis subunit; The fluctuation analysis subunit is used to compare the standard descent rate difference with the real-time maximum descent rate difference: If the difference between the real-time maximum descent rate and the standard descent rate is greater than the difference between the two, it is determined that the drying process is fluctuating drastically and a fault warning is issued. If the difference between the real-time maximum descent rate and the standard descent rate is less than or equal to the difference between the two rates, the drying process is considered stable and no fault warning is issued.

[0034] In this embodiment, the standard rate of decline difference represents the maximum allowable fluctuation limit of the rate of decline of image feature values ​​during an ideal and stable drying process. Based on historical high-quality production data, the standard rate of decline difference for the current seed variety is determined to be 0.02. By comparing the standard rate of decline difference with the real-time maximum rate of decline difference, the smoothness of the actual drying process is determined. If the actual fluctuation exceeds this threshold, the process is considered to be drastically fluctuating and poses a risk. For example, during the critical drying stage, sampling is performed every minute to obtain the following image feature value sequence: At time t0, the image feature value is 0.35; at time t1, the image feature value is 0.33; at time t2, the image feature value is 0.30; at time t3, the image feature value is 0.26; at time t4, the image feature value is 0.25; at time t5, the image feature value is 0.23; at time t6, the image feature value is 0.22. Calculate the decrease value per unit time (1 minute), i.e., the real-time decrease rate Vti, (i=1,2,……,6): Vt1=-0.02,Vt2=-0.03,Vt3=-0.04,Vt4=-0.01,Vt5=-0.02,Vt6=-0.01; Calculate the change between two adjacent real-time descent rates, i.e., the difference in descent rates ΔVj, (j=2, 3, ..., 6): ΔV2=-0.01, ΔV3=-0.01, ΔV4=+0.03, ΔV5=-0.01, ΔV6=+0.01; The process of determining the real-time maximum descent rate difference is as follows: find the maximum value among all |ΔV|, where |ΔV| are 0.01, 0.01, 0.03, 0.01, and 0.01 respectively. Therefore, the real-time maximum descent rate difference = 0.03. It is determined that the real-time maximum descent rate difference 0.03 is greater than the standard descent rate difference 0.02. The fault identification unit immediately triggers an early warning. At the same time, the first category identification subunit and the second category identification subunit are triggered to start in parallel to perform adhesion analysis and damage analysis respectively, and analyze the image to determine whether it is adhesion or damage.

[0035] Specifically, the category identification unit includes a first category identification subunit and a second category identification subunit; The first category identification subunit is used to extract the contour features of the particle outline based on the surface morphology image, determine whether there is particle adhesion based on the contour features, and output a warning prompt to increase the rotation speed or hot air temperature when it exists. The second category identification subunit is used to identify the damage to the surface of the pellets to be processed, so as to determine whether to output a warning prompt to reduce the rotation speed and hot air temperature.

[0036] Specifically, when the actual roundness is less than the standard roundness, it is judged as an abnormal contour shape; When the actual roundness is greater than or equal to the standard roundness, compare the actual aspect ratio with the standard aspect ratio: If the actual aspect ratio is less than or equal to the standard aspect ratio, it is judged as an abnormal contour shape; If the actual aspect ratio is greater than the standard aspect ratio, it is judged as a normal outline shape; Where, circularity = (4π × area) / (perimeter) 2 The actual aspect ratio is the ratio of the longer side to the shorter side of the smallest bounding rectangle.

[0037] In this embodiment, the ideal sphere has a roundness of 1, the standard roundness is 0.85, and the adhering clumps will be significantly smaller than 1; the standard aspect ratio is 1.5, and the adhering bodies are larger than 1.5.

[0038] In this embodiment, a first-category identification subunit is used to identify whether the pellets are stuck together, and a second-category identification subunit is used to identify whether the pellets are damaged on the surface, thus achieving precise control of the drying process. Foreground segmentation is performed on each frame of the image to extract all particle contours, and the contour features of each contour, including roundness and aspect ratio, are calculated. The contours of each particle are analyzed to obtain the proportion of abnormal particles, and this proportion is compared with a particle proportion threshold. If the proportion of abnormal particles is less than the particle proportion threshold, it is determined that there is no pellet adhesion; If the percentage of abnormal particles is greater than or equal to the particle percentage threshold, it is determined that there is pellet adhesion. The adjustment method is selected according to the degree of pellet adhesion, and an early warning prompt to increase the rotation speed or hot air temperature is output.

[0039] In this embodiment, the particle percentage threshold is set to 5%. When the abnormal particle percentage is greater than or equal to the particle percentage threshold, the specific process of selecting the adjustment method according to the degree of pellet adhesion is as follows: If 5% < abnormal particle percentage < 15%, the surface is judged to be slightly sticky, which may be due to slightly high local humidity, but the overall moisture evaporation trend is normal. However, it is still necessary to increase the rotation speed to enhance the rolling friction between the pellets and with the cylinder wall, and mechanically break up the slight adhesion, that is, increase the rotation speed of the rotary drum by 8%; if 15% ≤ abnormal particle percentage < 30%, it indicates that the surface humidity is too high or the temperature is insufficient, and the moisture evaporation is slow, resulting in sticky areas. To expand the drying capacity, the rotation speed needs to be increased, while the hot air temperature should be slightly increased. This dual approach of mechanical force and increased drying power will rapidly reduce surface humidity. Specifically, the rotary drum rotation speed should be increased by 15%, and the hot air set temperature by 3°C. If the percentage of abnormal particles is ≥30%, it indicates severe over-humidity or a formula problem, posing a risk of large-scale clumping. Conventional adjustments may be ineffective, requiring a significant increase in temperature and rotation speed, along with a warning to prevent the formation of dead clumps. The operator should be prompted to check the raw material humidity or coating formula. This involves increasing the hot air set temperature by 5-8°C and the rotation speed by 20%. Adjustments should be made dynamically based on trends to avoid over- or under-adjustment.

[0040] The parameter optimization module includes a polishing quality analysis unit and a dynamic adjustment unit; The polishing quality analysis unit is used to analyze the actual proportion of broken particles in the polishing stage based on the surface flow image, and compare the actual proportion of broken particles with the allowable proportion of broken particles. The dynamic adjustment unit is used to increase the absolute value of the standard descent rate of the next batch drying process in response to the actual proportion of broken particles being greater than the allowable proportion of broken particles, and adjust it to the corrected descent rate.

[0041] Specifically, the dynamic adjustment unit adjusts the standard descent rate according to the difference between the actual percentage of broken particles and the allowable percentage of broken particles, using the following formula: Corrected descent rate = standard descent rate × (1 + k × ΔR), where k is the adjustment coefficient, ΔR is the ratio of the actual percentage of broken particles exceeding the allowable percentage of broken particles to the allowable percentage of broken particles, k = 0.1. When ΔR is positive for multiple batches, the value of k is gradually reduced to achieve fine optimization.

[0042] In this embodiment, the percentage of broken particles is allowed to be set to 0.5%. The second image acquisition module is a high-speed industrial camera that captures a surface flow image of the moment of motion through the drum observation window, with the assistance of a flash at a specific frequency, for particle analysis. By segmenting the image into the foreground, all independent foreground contours are extracted, and feature analysis is performed on each contour to determine whether it is broken. The roundness and area of ​​the contour are calculated. For contours with acceptable roundness, their internal grayscale texture is further analyzed. Cracked particles will appear on the image. These cracks appear as continuous, dark linear textures running through the contour. Skeleton extraction or linear detection algorithms (such as Hough transform) can be used to identify them. The criteria are: intact pellets have a roundness > 0.85 and no internal crack features; cracked pellets have a roundness > 0.85 but obvious internal crack lines are detected; broken / scraped pellets have a roundness ≤ 0.7 or an area < 50% of the average area. The total number of cracked and broken / scraped pellets is counted as the total number of damaged contours. The total number of damaged contours is then calculated and compared with the total number of identified cracked / scraped pellets. The ratio of the number of pellet outlines yields the actual percentage of broken particles. This percentage is then compared to the allowable percentage to determine the quality of the current batch of polishing. If the actual percentage of broken particles is less than or equal to the allowable percentage, the batch is considered to be of acceptable quality. If the actual percentage is greater than the allowable percentage, the batch is considered to be of unacceptable quality. This damage is attributed to excessive microcracks caused by an overly vigorous drying process. The standard descent rate difference threshold for the next batch's drying process will be increased to ensure the next batch begins drying earlier. The process enters a slow-drying mode to reduce crack formation at the source. For example, if the total number of identified pellet outlines is 50, the number of cracked pellets is 3, and the number of broken / stripped pellets is 2, then the total number of damaged outlines is 5. The actual percentage of damaged particles is (5 / 50)×100%=10%>5%, indicating that the polishing quality of this batch is unqualified. The standard descent rate is adjusted from -0.03 to -0.033, converting the polishing effect into a measurable and feedback-able quantitative signal, thereby driving the continuous self-optimization of drying process parameters and ensuring the long-term stability of the process.

[0043] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent drying and polishing system for pelleting seeds, characterized by, The application relates to a drying and polishing device for processing pellets, which comprises the following parts: a rotating module for drying and polishing the pellets to be processed; a first image acquisition module connected with the rotating module, which is used for monitoring the surface topography image of the pellets to be processed in the drying stage, and determining the moisture characteristic curve of the pellets to be processed according to the surface topography image; a drying stage identification module connected with the first image acquisition module, which is used for identifying the drying stop node based on the moisture characteristic curve, or outputting a fault early warning prompt; a second image acquisition module connected with the rotating module, which is used for monitoring the surface flow state image of the pellets to be processed in the polishing stage; a parameter optimization module connected with the drying stage identification module and the second image acquisition module respectively, which is used for determining whether to adjust the standard falling rate of the next batch of drying process according to the analysis result of the surface flow state image.

2. The intelligent drying and polishing system for pelleting seeds as claimed in claim 1, wherein, The first image acquisition module comprises a shooting unit and a curve drawing unit; the shooting unit is used for periodically shooting the surface topography image of the pellets to be processed; the curve drawing unit is used for obtaining the surface glossiness and color saturation according to the surface topography image, calculating the image characteristic value according to the surface glossiness and color saturation, drawing the curve of the image characteristic value changing with time, and obtaining the moisture characteristic curve.

3. The intelligent drying and polishing system for pelleting seeds as claimed in claim 1, wherein, The drying stage identification module comprises a drying progress analysis unit, a fault identification unit and a category identification unit; the drying progress analysis unit is used for analyzing the moisture characteristic curve to obtain a first analysis result, a second analysis result and a third analysis result; the fault identification unit is used for obtaining the real-time maximum falling rate difference value of the moisture characteristic curve in response to the second analysis result, analyzing the fluctuation degree of the drying process according to the comparison result of the standard falling rate difference value and the real-time maximum falling rate difference value, and then determining whether to perform the fault early warning prompt; the category identification unit is used for identifying the fault cause and performing the fault early warning prompt in response to the real-time maximum falling rate difference value being greater than the standard falling rate difference value.

4. The intelligent drying and polishing system for pelleting seeds as claimed in claim 3, wherein, The drying stage identification module further comprises a node judgment unit; the node judgment unit is used for not performing the early warning prompt in response to the real-time maximum falling rate difference value being less than or equal to the standard falling rate difference value, and judging that the current time is the drying stop node until the absolute value of the real-time falling rate is less than or equal to the absolute value of the minimum falling rate.

5. The intelligent drying and polishing system for pelleting seeds as claimed in claim 3, wherein, The first analysis result is that the moisture characteristic curve does not have a peak value, and it is judged that the current drying stage is in a high moisture stage; the second analysis result is that the moisture characteristic curve has a peak value, and the absolute value of the real-time falling rate is greater than the absolute value of the standard falling rate, and it is judged that the drying critical stage is entered; the third analysis result is that the moisture characteristic curve has a peak value, and the absolute value of the real-time falling rate is less than or equal to the absolute value of the standard falling rate, and the hot air temperature is switched to a slow drying mode.

6. The intelligent drying and polishing system for pelleting seeds as claimed in claim 3, wherein, The fault identification unit comprises a falling rate calculation subunit; the falling rate calculation subunit is used for calculating a plurality of real-time falling rate difference values and the real-time maximum falling rate difference value among the real-time falling rate difference values based on the moisture characteristic curve. The moisture feature curve is obtained, and the image feature values of each sampling point on the moisture feature curve are obtained for a current sampling period.

7. The intelligent drying and polishing system for pelleting seeds according to claim 6, characterized in that, The fault recognition unit further comprises a fluctuation analysis subunit; The fluctuation analysis subunit is configured to compare the standard drop rate difference with the real-time maximum drop rate difference: If the real-time maximum drop rate difference is greater than the standard drop rate difference, it is determined that the drying process is in a violent fluctuation, and a fault warning is prompted; If the real-time maximum drop rate difference is less than or equal to the standard drop rate difference, it is determined that the drying process is stable, and no fault warning is prompted.

8. The intelligent drying and polishing system for pelleting seeds as claimed in claim 3, wherein, The category recognition unit comprises a first category recognition subunit and a second category recognition subunit. The first category recognition subunit is configured to extract contour features of a particle contour based on the surface topography image, determine whether there is agglomeration of pellets based on the contour features, and output a pre-warning prompt for increasing the rotation speed or the hot air temperature when there is agglomeration of pellets. The second category recognition subunit is configured to identify a damage condition of a surface of a pellet to be processed, and determine whether to output a pre-warning prompt for reducing the rotation speed and the hot air temperature.

9. The intelligent drying and polishing system for pelleting seeds as claimed in claim 1, wherein, The parameter optimization module comprises a polishing quality analysis unit and a dynamic adjustment unit. The polishing quality analysis unit is configured to analyze an actual damaged particle proportion in a polishing stage, and compare the actual damaged particle proportion with an allowable damaged particle proportion. The dynamic adjustment unit is configured to increase an absolute value of the standard drop rate of a next batch of drying procedures in response to the actual damaged particle proportion being greater than the allowable damaged particle proportion.

10. The intelligent drying and polishing system for pelleting seeds as claimed in claim 1, wherein, The rotation module comprises: A rotary cylinder body having an inner wall provided with switchable scoops for lifting materials during drying and smooth lining plates during polishing; A driving unit configured to drive the rotary cylinder body to rotate; A hot air supply unit configured to control the temperature, humidity, and air volume of hot air in the rotary cylinder body.

Images