A method and system for processing feed pellets using fermented silkworm excrement as raw material

Abstract

This invention belongs to the field of anomaly identification technology and proposes a method and system for processing feed pellets using fermented silkworm excrement as raw material. Specifically, an infrared thermal imager is first placed in the conditioner of the feed pellet mill. During the low-temperature pelleting process, infrared images are acquired and stored in real time using the infrared thermal imager to obtain an infrared image sequence. Then, the infrared images are divided into regions to obtain several processing zones. Finally, a gelatinization loss risk analysis is performed on the low-temperature processing of each micro-processing zone to obtain gelatinization null values. Based on these null values, gelatinization anomalies are identified in the feed. By using gelatinization loss risk analysis, the invention accurately addresses localized gelatinization lag caused by microscopic mechanisms such as delayed heat transfer paths, differences in particle heat capacity, and limited moisture migration. It locates feed pellets at risk of pulverization in the low-temperature processed feed, thereby reducing the risk of pulverization in feed made from fermented silkworm excrement and improving feed quality.

Description

Technical Field

[0001] This invention belongs to the field of anomaly identification technology, specifically relating to a method and system for processing feed pellets using fermented silkworm excrement as raw material. Background Technology

[0002] Fermented silkworm excrement, as a feed ingredient, can significantly promote animal growth. This growth promotion is mainly due to the effect of fermented silkworm excrement on regulating the intestinal microecology. Fermented silkworm excrement is rich in probiotics, typically including lactic acid bacteria and yeast. These beneficial bacteria greatly promote the balance of the intestinal microbiota, improve the immunity of the animal's digestive system, and reduce the risk of intestinal diseases. Simultaneously, fermented silkworm excrement can improve feed utilization. Silkworm excrement is rich in crude fiber, and the cellulose and hemicellulose in silkworm excrement and basic feed are inherently difficult to digest. However, microbial fermentation can degrade these components, making them easier for animals to absorb. One method of producing feed from fermented silkworm excrement is to mix it with other feed ingredients and then extrude it into pellets using a pellet mill, improving palatability and storage stability. However, the pelleting process requires high temperatures to gelatinize raw materials such as corn and rice bran, making the pellets more viscous and reducing the risk of pulverization. But high temperatures inevitably lead to the death of a large number of probiotics in the fermented silkworm excrement. Therefore, a low-temperature pelleting method is more suitable. The most direct drawback of the low-temperature pelleting method is the pulverization problem caused by insufficient gelatinization. Therefore, there is an urgent need for a method and system for processing feed pellets using fermented silkworm excrement as raw material. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for processing feed pellets using fermented silkworm excrement as raw material, in order to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0004] To achieve the above objectives, according to one aspect of the present invention, a method for processing feed pellets using fermented silkworm excrement as raw material is provided, the method comprising the following steps:

[0005] S100 is an infrared thermal imager equipped in a feed pellet drying device.

[0006] S200 uses an infrared thermal imager to collect and store infrared images in real time during the low-temperature granulation process, thus obtaining an infrared image sequence.

[0007] S300, divides the infrared image into regions to obtain several processing areas;

[0008] S400 performs a gelatinization loss risk analysis on the low-temperature processing of each microprocessor zone to obtain gelatinization null values.

[0009] S500 identifies gelatinization anomalies in feed based on gelatinization null values.

[0010] Further, in step S100, the method for equipping the feed pellet drying device with an infrared thermal imager is as follows: the feed pellet drying device includes a low-temperature drying chamber, a hot air drying chamber, or a low-temperature air drying chamber; an infrared thermal imager is arranged in the feed pellet drying device so that the infrared thermal imager is positioned at a visible position of the surface heat radiation of the feed pellet forming area.

[0011] Infrared thermal imagers can also be infrared scanners, near-infrared industrial cameras, or multispectral imagers. The structural designs of different feed pellet mill models may vary significantly. If installation conditions are lacking, such as space constraints or insufficient electromagnetic interference protection, hardware adaptation via custom brackets or adapter modules is required.

[0012] In traditional feed pellet drying equipment, the low-temperature pelleting strategy leads to uneven temperature distribution in the forming zone, easily resulting in under-gelatinized pellets in areas with insufficient heat conduction, thus causing pulverization. This method utilizes an infrared thermal imaging device installed in the pelleting channel or pellet extrusion outlet of the feed pelleting equipment to acquire real-time images of the thermal radiation on the surface of the pellets formed by extrusion of feed raw materials during the pelleting process. The infrared thermal imaging device captures the temperature distribution in various local areas during pelleting through non-contact infrared sensing, and the acquired image data will subsequently serve as the basis for determining the presence of gelatinized voids.

[0013] Further, in step S200, during the low-temperature granulation process, infrared images are acquired and stored in real time using an infrared thermal imager to obtain an infrared image sequence. The method is as follows: when the low-temperature drying oven starts granulation, the infrared thermal imager acquires and records the thermal radiation images inside the oven in real time until the low-temperature granulation process ends, forming an infrared image sequence from all the images. The acquisition interval is 5-20 seconds; that is, the acquired images are stored according to a set period to form an infrared image sequence of feed pellets during the drying process. The low-temperature granulation process lasts from 2 to 24 hours, and the total number of images in the infrared image sequence should be at least 500 to fully explore microscopic change trends.

[0014] Infrared images reflect the temperature distribution on the particle surface at any drying moment. Continuously acquired infrared images constitute the temporal temperature evolution trajectory of the drying process, which is used for the gelatinization risk assessment of each subdivided region in subsequent steps. It is the core data foundation for subsequent regional heat treatment optimization control.

[0015] Furthermore, in step S300, the method for dividing the infrared image into regions to obtain several processing areas is as follows: each frame of the infrared image sequence is spatially divided into regularized grids to obtain an image grid, which is denoted as a microprocessing area, and each microprocessing area is associated with a predetermined area in the drying oven.

[0016] Regularized spatial partitioning refers to processing infrared thermal imaging images into a two-dimensional grid according to uniform geometric rules to achieve a structured representation of the image space. This is achieved through the `np.zeros()` function in OpenCV. Regularized partitioning is a fixed grid slicing algorithm based on the image pixel matrix. It divides the original thermal image into m×n rectangular processing areas, which are called microprocessing areas. Each microprocessing area corresponds to a set of pixels in a region of the image.

[0017] Further, in step S400, the method for obtaining the null value by performing a gelatinization loss risk analysis on the low-temperature processing of each microprocessor area is as follows: the median value of the pixel value of the microprocessor area is recorded as its corresponding gray value, which constitutes the gray value time series of the microprocessor area in the infrared image sequence; the gray value time series is subjected to first-order difference, and the gray change rate of each frame is extracted.

[0018] The first-order difference refers to subtracting any element from the first element in the reverse time direction, and the ratio of the difference to the first element in the reverse time direction is the grayscale change rate.

[0019] A continuous frame segment with a change rate amplitude lower than a preset fluctuation threshold is identified as a plateau period, and the number of continuous frames of this segment is extracted as the plateau period length Tf;

[0020] The preset fluctuation threshold is ε∈[0.3%,1.5%]. If the thermal inertia of the heating process is large, it is recommended to set a smaller fluctuation threshold. A large thermal inertia of the heating process can also be equivalent to the heating rate of the heating process. Alternatively, if the infrared image noise is low, the preset fluctuation threshold can be set to a small threshold. A continuous frame segment is composed of each consecutive moment where the magnitude of the coincidence rate of change is lower than the preset fluctuation threshold. Each moment where the magnitude of the coincidence rate of change is lower than the preset fluctuation threshold belongs to only one continuous frame segment.

[0021] Extract the starting time position Ts of the plateau period; where the starting time position is the sequence number of the frame in the grayscale time series;

[0022] The null values ​​of the microprocessor region are obtained by constructing a weighted function based on Tf and Ts.

[0023] The specific process is as follows: the ratio of the plateau period length to the total length of the grayscale value time series is used as the participation weight Pwt, and the ratio of the starting time position to the total length of the grayscale value time series is used as the trigger position weight Tpwt. Then the null value is Gnv=Pwt×exp(﹣α×Tpwt); where α is the preset plateau period concern coefficient, and the value range is [0.3,1].

[0024] The shorter the plateau segment, the greater the need to improve the sensitivity to gelatinization null values, so a larger plateau period concern coefficient should be used; otherwise, a smaller value should be used, with a default value of 1. Under the same feed material ratio, the same value can be used for the plateau period concern coefficient.

[0025] The gelatinization null value calculated at time points quantifies the response time difference between initial heating and the formation of the thermal equilibrium zone, and uses plateau stability as a weighted correction. It microscopically maps the degree of thermal response lag caused by differences in regional heat capacity, fiber distribution, or moisture content. Since it is a measurement method based on time series feature extraction, it does not rely on the structure and material of the pellet mill or drying device for direct detection. It effectively explores the reproducibility and quantifiability of thermal state evolution in the time dimension, and thus provides a mathematical basis for further realizing the fine identification or compensation measures for the internal thermal field uniformity of complex pelleting systems. However, the aforementioned blurred null values ​​constructed based on the starting frame and duration of the plateau segment, while quantifying the trend of thermal response lag within the region in the time dimension, still have two key limitations: Firstly, they rely excessively on the accurate identification of the plateau segment. When the image grayscale signal fluctuates significantly or the plateau segment itself is short, the plateau boundary is easily affected by noise, leading to increased volatility in the blurred null values. Secondly, this process relies on overall thermal stability but fails to further characterize the clarity and persistence of the spatial distribution of abnormal regions, making it difficult to effectively identify high-risk areas with significant local thermal lag. Therefore, this method also provides another method for calculating blurred null values ​​as follows:

[0026] Further, in step S400, the method for performing a gelatinization loss risk analysis on the low-temperature processing of each microprocessor area to obtain gelatinization null values ​​is as follows: the lower quartile value of each pixel value in the microprocessor area is recorded as the inefficient marker value, and the difference between the inefficient marker value and the minimum pixel value in the microprocessor area is the inefficient interval; the difference between the inefficient marker value at any time and the previous time is defined as the inefficient marker increment, and if the value of the inefficient marker increment is negative, an inefficient event is marked at the corresponding time, and the ratio of the number of times when an inefficient event occurs to the number of times when no inefficient event occurs is recorded as the inefficient coefficient Tsic;

[0027] The ratio of the inefficient interval to the inefficient marker value is denoted as the inefficient product Iev. The median of the inefficient product at the moment an inefficient event occurs is the product level Iev.base. If the inefficient product corresponding to an inefficient event is less than the product level, that moment is marked as a preheating void. If the inefficient product corresponding to a preheating void is smaller than the inefficient product corresponding to the moment of the first inefficient event in the reverse time direction, that moment is defined as the first cavitation void. For any first cavitation void, the moment of the first inefficient event with an inefficient product higher than the product level is searched in the forward time direction. The number of moments between the obtained moment and any first cavitation void is denoted as the cavitation recovery distance Crd. The cavitation void value Gnv is calculated based on the cavitation recovery distance of the first cavitation void and the transverse inefficiency coefficient. Its mathematical expression is as follows:

[0028] ;

[0029] Where i1 is the cumulative variable, lg() is the logarithmic function with base 10, exp() is the exponential function with base e and natural constant e, and Crd i1 and Iev i1 represents the cavitation recovery distance and inefficient product corresponding to the i1th first cavitation void, respectively, and num is the number of first cavitation voids.

[0030] The previous moment refers to the moment in the reverse time direction; the normalization process uses the minmax method to limit all values ​​to between 0 and 1.

[0031] Beneficial effects: By modeling the time-series changes of the grayscale values ​​of infrared images in the processing area during the drying process, behavioral features reflecting heat absorption lag or abnormal temperature response are extracted to calculate the gelatinization null value. This enables a quantitative assessment of the risk of gelatinization loss at a local spatial scale. It quantifies the gelatinization lag behavior caused by internal heat transfer barriers, differences in heat capacity, or differences in moisture migration rates within the material, thus providing data support for subsequent abnormal area identification and differentiated heat supplementation intervention. This makes the pelleting process more dynamic and responsive, significantly improving the consistency and controllability of feed pellets.

[0032] Further, in step S500, the method for identifying gelatinization anomalies in feed based on gelatinization null values ​​is as follows: the gelatinization null values ​​corresponding to each microprocessor region are used to form a gelatinization null value set and are standardized. The upper quartile value and standard deviation of the null value set are recorded as the first anomaly threshold and the threshold gradient, respectively. The microprocessor region corresponding to the element in the null value set that is greater than the upper quartile value is recorded as the first anomaly region. The difference between the gelatinization null value in the first anomaly region and the threshold gradient is recorded as the second anomaly threshold. If the set of gelatinization null values ​​corresponding to the eight neighbors of the first anomaly region is recorded as the gelatinization gradient set, and if more than half of the elements in the gelatinization gradient set are greater than the second anomaly threshold, then it is determined that a gelatinization anomaly has occurred in the first anomaly region; otherwise, no gelatinization anomaly has occurred.

[0033] This step identifies anomalies based on the distribution patterns of the gelatinization null value set and local spatial thermal diffusion behavior. It characterizes globally high-value areas by introducing the upper quartile and standard deviation, and extracts local gradient responses by combining the eight-neighborhood gelatinization null value change trends. This maps the continuity principle and minimum heat loss path principle of the heat conduction process from a microscopic perspective. Because heat diffusion tends to flow towards areas with larger temperature gradients, if the gelatinization null value in a certain processing area is significantly higher than the surrounding area, and the surrounding area also shows a synchronous upward trend, it indicates energy stagnation or material thermal resistance imbalance, violating the trend of uniform heat conduction and suggesting local gelatinization obstacles in actual granulation. Therefore, this method achieves accurate gelatinization anomaly determination from two dimensions: data distribution statistics and spatial thermal field perturbation.

[0034] Furthermore, in step S500, the method for identifying gelatinization abnormalities in feed based on gelatinization null values ​​also includes: screening the feed particles corresponding to the microprocessing area where gelatinization abnormalities occur, and then performing high-temperature steam gelatinization treatment.

[0035] Low-temperature pelleting aims to preserve the activity of probiotics such as lactic acid bacteria and yeast in fermented silkworm excrement. However, it can easily lead to insufficient local heat, resulting in problems such as incomplete pellet gelatinization, loose structure, or increased pulverization rate. Applying high-temperature heating to the entire batch of feed at once would cause widespread inactivation of probiotics, contradicting the original purpose of low-temperature pelleting. Therefore, this invention uses image recognition to accurately locate abnormal areas and performs localized high-temperature steam gelatinization only on the identified areas of gelatinization abnormality. This ensures pellet quality while maximizing the preservation of overall probiotic activity, achieving a dual guarantee of feed functionality and structural performance.

[0036] Preferably, all undefined variables in this invention, if not explicitly defined, can be manually set thresholds.

[0037] This invention also provides a feed pellet processing system using fermented silkworm excrement as raw material. The system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the feed pellet processing method using fermented silkworm excrement as raw material. This feed pellet processing system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. The runnable system may include, but is not limited to, processors, memory, and server clusters. The processor executes the computer program within the following system units:

[0038] Infrared thermal imaging arrangement unit, used to arrange infrared thermal imagers in the conditioner of a feed pellet mill;

[0039] The real-time monitoring unit is used to acquire and store infrared images in real time during the low-temperature granulation process using an infrared thermal imager, thereby obtaining an infrared image sequence.

[0040] The region division unit is used to divide the infrared image into regions to obtain several processing areas;

[0041] The gelatinization loss analysis unit is used to perform gelatinization loss risk analysis on the low-temperature processing of each microprocessor zone to obtain gelatinization null values.

[0042] The gelatinization anomaly identification unit is used to identify gelatinization anomalies in feed based on gelatinization null values.

[0043] The beneficial effects of this invention are as follows: Based on the sluggish thermal response characteristics of fermented silkworm excrement under low-temperature pelleting conditions, and considering its high cellulose and hemicellulose content, uneven moisture migration channels, and complex heat capacity distribution, a method for identifying feed gelatinization deficiency is proposed. By using gelatinization null value quantification, the degree of localized insufficient gelatinization caused by factors such as sluggish heat conduction, differences in heat capacity, or limited moisture migration within the pellets is objectively characterized through image thermal response data during the low-temperature pelleting process. This effectively solves the problem of localized pulverization due to insufficient overall heat while protecting the activity of probiotics in fermented silkworm excrement during low-temperature pelleting. It overcomes the limitation of traditional processes where gelatinization uniformity and microbial activity protection are difficult to achieve simultaneously, significantly improving feed quality factors such as the structural stability and storage palatability of feed pellets. Attached Figure Description

[0044] The above and other features of the present invention will become more apparent from the detailed description of the embodiments shown in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings:

[0045] Figure 1 The diagram shows a flowchart of a method for processing feed pellets using fermented silkworm excrement as raw material.

[0046] Figure 2 The diagram shows a structural diagram of a feed pellet processing system using fermented silkworm excrement as raw material. Detailed Implementation

[0047] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0048] Example 1

[0049] like Figure 1 The diagram shows a flowchart of a method for processing feed pellets using fermented silkworm excrement as raw material. The following section will discuss this method in conjunction with... Figure 1 This invention describes a method for processing feed pellets using fermented silkworm excrement as raw material, comprising the following steps:

[0050] S100 is an infrared thermal imager equipped in a feed pellet drying device.

[0051] S200 uses an infrared thermal imager to collect and store infrared images in real time during the low-temperature granulation process, thus obtaining an infrared image sequence.

[0052] S300, divides the infrared image into regions to obtain several processing areas;

[0053] S400 performs a gelatinization loss risk analysis on the low-temperature processing of each microprocessor zone to obtain gelatinization null values.

[0054] S500 identifies gelatinization anomalies in feed based on gelatinization null values.

[0055] Further, in step S100, the method for equipping the feed pellet drying device with an infrared thermal imager is as follows: the feed pellet drying device includes a low-temperature drying chamber, a hot air drying chamber, or a low-temperature air drying chamber; an infrared thermal imager is arranged in the feed pellet drying device so that the infrared thermal imager is positioned at a visible position of the surface heat radiation of the feed pellet forming area.

[0056] Infrared thermal imagers can also be infrared scanners, near-infrared industrial cameras, or multispectral imagers. The structural designs of different feed pellet mill models may vary significantly. If installation conditions are lacking, such as space constraints or insufficient electromagnetic interference protection, hardware adaptation via custom brackets or adapter modules is required.

[0057] Further, in step S200, during the low-temperature granulation process, infrared images are acquired and stored in real time using an infrared thermal imager to obtain an infrared image sequence. The method is as follows: when the low-temperature drying oven starts granulation, the infrared thermal imager acquires and records the thermal radiation images inside the oven in real time until the low-temperature granulation process ends, forming an infrared image sequence from all the images. The acquisition interval is 5-20 seconds; that is, the acquired images are stored according to a set period to form an infrared image sequence of feed pellets during the drying process. The low-temperature granulation process lasts from 2 to 24 hours, and the total number of images in the infrared image sequence should be at least 500 to fully explore microscopic change trends.

[0058] Furthermore, in step S300, the method for dividing the infrared image into regions to obtain several processing areas is as follows: each frame of the infrared image sequence is spatially divided into regularized grids to obtain an image grid, which is denoted as a microprocessing area, and each microprocessing area is associated with a predetermined area in the drying oven.

[0059] Regularized spatial partitioning refers to processing infrared thermal imaging images into a two-dimensional grid according to uniform geometric rules to achieve a structured representation of the image space. This is achieved through the `np.zeros()` function in OpenCV. Regularized partitioning is a fixed grid slicing algorithm based on the image pixel matrix. It divides the original thermal image into m×n rectangular processing areas, which are called microprocessing areas. Each microprocessing area corresponds to a set of pixels in a region of the image.

[0060] Further, in step S400, the method for obtaining the null value by performing a gelatinization loss risk analysis on the low-temperature processing of each microprocessor area is as follows: the median value of the pixel value of the microprocessor area is recorded as its corresponding gray value, which constitutes the gray value time series of the microprocessor area in the infrared image sequence; the gray value time series is subjected to first-order difference, and the gray change rate of each frame is extracted.

[0061] The first-order difference refers to subtracting any element from the first element in the reverse time direction, and the ratio of the difference to the first element in the reverse time direction is the grayscale change rate.

[0062] A continuous frame segment with a change rate amplitude lower than a preset fluctuation threshold is identified as a plateau period, and the number of continuous frames of this segment is extracted as the plateau period length Tf;

[0063] The preset fluctuation threshold is ε∈[0.3%,1.5%]. If the thermal inertia of the heating process is large, it is recommended to set a smaller fluctuation threshold. A large thermal inertia of the heating process can also be equivalent to the heating rate of the heating process. Alternatively, if the infrared image noise is low, the preset fluctuation threshold can be set to a small threshold. A continuous frame segment is composed of each consecutive moment where the magnitude of the coincidence rate of change is lower than the preset fluctuation threshold. Each moment where the magnitude of the coincidence rate of change is lower than the preset fluctuation threshold belongs to only one continuous frame segment.

[0064] Extract the starting time position Ts of the plateau period; where the starting time position is the sequence number of the frame in the grayscale time series;

[0065] The null values ​​of the microprocessor region are obtained by constructing a weighted function based on Tf and Ts.

[0066] The specific process is as follows: the ratio of the plateau period length to the total length of the grayscale value time series is used as the participation weight Pwt, and the ratio of the starting time position to the total length of the grayscale value time series is used as the trigger position weight Tpwt. Then the null value is Gnv=Pwt×exp(﹣α×Tpwt); where α is the preset plateau period concern coefficient, and the value range is [0.3,1].

[0067] The shorter the plateau segment, the greater the need to improve the sensitivity to gelatinization null values, so a larger plateau period concern coefficient should be used; otherwise, a smaller value should be used, with a default value of 1. Under the same feed material ratio, the same value can be used for the plateau period concern coefficient.

[0068] Further, in step S500, the method for identifying gelatinization anomalies in feed based on gelatinization null values ​​is as follows: the gelatinization null values ​​corresponding to each microprocessor region are used to form a gelatinization null value set and are standardized. The upper quartile value and standard deviation of the null value set are recorded as the first anomaly threshold and the threshold gradient, respectively. The microprocessor region corresponding to the element in the null value set that is greater than the upper quartile value is recorded as the first anomaly region. The difference between the gelatinization null value in the first anomaly region and the threshold gradient is recorded as the second anomaly threshold. If the set of gelatinization null values ​​corresponding to the eight neighbors of the first anomaly region is recorded as the gelatinization gradient set, and if more than half of the elements in the gelatinization gradient set are greater than the second anomaly threshold, then it is determined that a gelatinization anomaly has occurred in the first anomaly region; otherwise, no gelatinization anomaly has occurred.

[0069] This step identifies anomalies based on the distribution patterns of the gelatinization null value set and local spatial thermal diffusion behavior. It characterizes globally high-value areas by introducing the upper quartile and standard deviation, and extracts local gradient responses by combining the eight-neighborhood gelatinization null value change trends. This maps the continuity principle and minimum heat loss path principle of the heat conduction process from a microscopic perspective. Because heat diffusion tends to flow towards areas with larger temperature gradients, if the gelatinization null value in a certain processing area is significantly higher than the surrounding area, and the surrounding area also shows a synchronous upward trend, it indicates energy stagnation or material thermal resistance imbalance, violating the trend of uniform heat conduction and suggesting local gelatinization obstacles in actual granulation. Therefore, this method achieves accurate gelatinization anomaly determination from two dimensions: data distribution statistics and spatial thermal field perturbation.

[0070] Furthermore, in step S500, the method for identifying gelatinization abnormalities in feed based on gelatinization null values ​​also includes: screening the feed particles corresponding to the microprocessing area where gelatinization abnormalities occur, and then performing high-temperature steam gelatinization treatment.

[0071] Low-temperature pelleting aims to preserve the activity of probiotics such as lactic acid bacteria and yeast in fermented silkworm excrement. However, it can easily lead to insufficient local heat, resulting in problems such as incomplete pellet gelatinization, loose structure, or increased pulverization rate. Applying high-temperature heating to the entire batch of feed at once would cause widespread inactivation of probiotics, contradicting the original purpose of low-temperature pelleting. Therefore, this invention uses image recognition to accurately locate abnormal areas and performs localized high-temperature steam gelatinization only on the identified areas of gelatinization abnormality. This ensures pellet quality while maximizing the preservation of overall probiotic activity, achieving a dual guarantee of feed functionality and structural performance.

[0072] Example 2

[0073] Example 2 uses the same gelatinization anomaly identification method as Example 1. The difference is that in step S400, the method for obtaining gelatinization null values ​​by performing gelatinization loss risk analysis on the low-temperature processing of each microprocessor area is as follows: the lower quartile value of each pixel value in the microprocessor area is recorded as the inefficient marker value, and the difference between the inefficient marker value and the minimum pixel value in the microprocessor area is the inefficient interval; the difference between the inefficient marker value at any time and the previous time is defined as the inefficient marker increment. If the value of the inefficient marker increment is negative, an inefficient event is marked at the corresponding time. The ratio of the number of times an inefficient event occurs to the number of times an inefficient event does not occur is recorded as the inefficient coefficient Tsic.

[0074] The ratio of the inefficient interval to the inefficient marker value is denoted as the inefficient product Iev. The median of the inefficient product at the moment an inefficient event occurs is the product level Iev.base. If the inefficient product corresponding to an inefficient event is less than the product level, that moment is marked as a preheating void. If the inefficient product corresponding to a preheating void is smaller than the inefficient product corresponding to the moment of the first inefficient event in the reverse time direction, that moment is defined as the first cavitation void. For any first cavitation void, the moment of the first inefficient event with an inefficient product higher than the product level is searched in the forward time direction. The number of moments between the obtained moment and any first cavitation void is denoted as the cavitation recovery distance Crd. The cavitation void value Gnv is calculated based on the cavitation recovery distance of the first cavitation void and the transverse inefficiency coefficient. Its mathematical expression is as follows:

[0075] ;

[0076] Where i1 is the cumulative variable, lg() is the logarithmic function with base 10, exp() is the exponential function with base e and natural constant e, and Crd i1 and Iev i1 represents the cavitation recovery distance and inefficient product corresponding to the i1th first cavitation void, respectively, and num is the number of first cavitation voids.

[0077] The previous moment refers to the moment in the reverse time direction; the normalization process uses the minmax method to limit all values ​​to between 0 and 1.

[0078] To verify the effectiveness of the proposed gelatinization anomaly identification and heat replenishment method in alleviating pulverization problems during low-temperature granulation, the following experimental scheme was designed, and the following comparative data were obtained. The same batch of fermented silkworm excrement feed raw materials were selected, and three control experiments were set up: a comparative example, Example 1, and Example 2. The comparative example used only the traditional low-temperature granulation method, without introducing any infrared image recognition mechanism or subsequent heat replenishment mechanism. Example 1 identified gelatinization anomalies in the microprocessing area based on the first-order difference and fluctuation threshold of the image plateau period. Example 2 evaluated anomalies based on the plateau grayscale change rate and system judgment coefficient. After granulation and sieving, the weights of the identified abnormal areas, correctly identified pulverized areas, and misidentified areas were measured for the three groups of feed pellets, and the pulverization identification accuracy rate was calculated accordingly. In Examples 1 and 2, half of the weight of the identified abnormal areas was used for the subsequent heat replenishment mechanism, and the weight was extrapolated back to the complete processed weight to obtain the differentiation rate after heat replenishment.

[0079] Table 1. Comparison of Powdering Anomaly Detection Rates

[0080]

[0081] As shown in Table 1, under the same batch of feed ingredients and low-temperature pelleting conditions, both Example 1 and Example 2 demonstrated significantly better ability to identify powdering anomalies than the comparative example. Example 2, by constructing a combination mechanism between the grayscale change rate during the plateau period and the judgment coefficient, achieved a powdering identification accuracy rate of 93.5%, significantly higher than Example 1's 89.7% and the comparative example's 0%, indicating greater sensitivity and accuracy in judging microscopic image grayscale differences. In contrast, the comparative example, without using infrared image analysis, could not actively identify potential powdering areas, leading to the inability to perform subsequent high-temperature supplementary heating and resulting in insufficient feed quality. This method, through precise identification of the gelatinization lag area, significantly improves the targeting and efficiency of supplementary heating treatment, maximizing the protection of probiotic activity and ensuring pellet stability, thus verifying the technical superiority and practical value of this invention in the field of animal feed pelleting.

[0082] An embodiment of the present invention provides a feed pellet processing system using fermented silkworm excrement as raw material, such as... Figure 2 The diagram shows a structural diagram of a feed pellet processing system using fermented silkworm excrement as raw material according to the present invention. The feed pellet processing system using fermented silkworm excrement as raw material in this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the above-described embodiment of the feed pellet processing method using fermented silkworm excrement as raw material.

[0083] The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program in units of the following system:

[0084] Infrared thermal imaging arrangement unit, used to arrange infrared thermal imagers in the conditioner of a feed pellet mill;

[0085] The real-time monitoring unit is used to acquire and store infrared images in real time during the low-temperature granulation process using an infrared thermal imager, thereby obtaining an infrared image sequence.

[0086] The region division unit is used to divide the infrared image into regions to obtain several processing areas;

[0087] The gelatinization loss analysis unit is used to perform gelatinization loss risk analysis on the low-temperature processing of each microprocessor zone to obtain gelatinization null values.

[0088] The gelatinization anomaly identification unit is used to identify gelatinization anomalies in feed based on gelatinization null values.

[0089] The feed pellet processing system using fermented silkworm excrement as raw material can run on computing devices such as desktop computers, laptops, handheld computers, and cloud servers. The system that can run on the feed pellet processing system using fermented silkworm excrement as raw material may include, but is not limited to, processors and memory. Those skilled in the art will understand that the example described is merely an illustration of a feed pellet processing system using fermented silkworm excrement as raw material and does not constitute a limitation on such a system. It may include more or fewer components, or a combination of certain components, or different components. For example, the feed pellet processing system using fermented silkworm excrement as raw material may also include input / output devices, network access devices, buses, etc.

[0090] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the feed pellet processing system using fermented silkworm excrement as raw material, connecting all parts of the system via various interfaces and lines.

[0091] The memory can be used to store the computer programs and / or modules. The processor, by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory, realizes various functions of the feed pellet processing system using fermented silkworm excrement as raw material. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0092] Although the invention has been described in considerable detail and particularly with regard to several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, thereby effectively covering the intended scope of the invention. Furthermore, the invention has been described above with respect to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for processing feed pellets using fermented silkworm excrement as raw material, characterized in that, The method includes the following steps: S100 is an infrared thermal imager equipped in a feed pellet drying device. S200 uses an infrared thermal imager to collect and store infrared images in real time during the low-temperature granulation process, thus obtaining an infrared image sequence. S300 divides the infrared image into regions to obtain several micro-processing areas; S400 performs a gelatinization loss risk analysis on the low-temperature processing of each microprocessor zone to obtain gelatinization null values. S500 identifies gelatinization anomalies in feed based on gelatinization null values. In step S300, the method for dividing the infrared image into regions to obtain several microprocessor areas is as follows: each frame of the infrared image in the infrared image sequence is divided into regular spatial regions to obtain an image grid, which is recorded as a microprocessor area, so that each microprocessor area is associated with a predetermined area in the drying oven. Step S400 involves performing a gelatinization loss risk analysis on the low-temperature processing of each microprocessor region to obtain gelatinization null values. The method is as follows: The median value of each pixel in the microprocessor region is recorded as its corresponding grayscale value, forming a grayscale value time series of the microprocessor region in the infrared image sequence. First-order difference is performed on the grayscale value time series to extract the grayscale change rate of each frame. A continuous frame segment with a change rate amplitude lower than a preset fluctuation threshold is identified as a plateau period, and the number of consecutive frames in this segment is extracted as the plateau period length Tf. The starting time position Ts of the plateau period is extracted. A weighted function is constructed based on Tf and Ts to obtain the gelatinization null values ​​of the microprocessor region. Alternatively, step S400 can be used to perform a gelatinization loss risk analysis on the low-temperature processing of each microprocessor area to obtain gelatinization null values. The lower quartile value of each pixel value in the microprocessor area is recorded as the inefficient marker value, and the difference between the inefficient marker value and the minimum pixel value in the microprocessor area is the inefficient interval. The difference between the inefficient marker value at any given time and the inefficient marker value at the previous time is defined as the inefficient marker increment. If the value of the inefficient marker increment is negative, an inefficient event is marked at the corresponding time. The ratio of the number of times an inefficient event occurs to the number of times an inefficient event does not occur is recorded as the inefficient coefficient Tsic. The ratio of the inefficient interval to the inefficient marker value is denoted as the inefficient product. The median of the inefficient product at the moment an inefficient event occurs is the product level. If the inefficient product corresponding to an inefficient event is less than the product level, that moment is marked as a preheating void. If the inefficient product value corresponding to the preheating void is smaller than that corresponding to the moment of the first inefficient event in the reverse time direction, that moment is defined as the first cavitation void. For any first cavitation void, search for the moment of the first inefficient event with an inefficient product higher than the product level in the forward time direction. The number of moments between the obtained moment and any first cavitation void is denoted as the cavitation recovery distance. The cavitation void value is calculated based on the cavitation recovery distance of the first cavitation void and the lateral inefficiency coefficient.

2. The method for processing feed pellets using fermented silkworm excrement as raw material according to claim 1, characterized in that, In step S100, the method for equipping the feed pellet drying device with an infrared thermal imager is as follows: the feed pellet drying device includes a low-temperature drying chamber, a hot air drying chamber, or a low-temperature air drying chamber; an infrared thermal imager is arranged in the feed pellet drying device so that the infrared thermal imager is positioned at a visible position of the surface heat radiation of the feed pellet forming area.

3. The method for processing feed pellets using fermented silkworm excrement as raw material according to claim 1, characterized in that, In step S200, during the low-temperature granulation process, infrared images are acquired and stored in real time using an infrared thermal imager to obtain an infrared image sequence. The method is as follows: when the low-temperature drying oven starts low-temperature granulation, the infrared thermal imager acquires thermal radiation images inside the low-temperature drying oven in real time and records them as infrared images until the low-temperature granulation process ends, and all infrared images are used to form an infrared image sequence.

4. The method for processing feed pellets using fermented silkworm excrement as raw material according to claim 1, characterized in that, In step S500, the method for identifying gelatinization anomalies in feed based on gelatinization null values ​​is as follows: the gelatinization null values ​​corresponding to each microprocessor region are used to form a gelatinization null value set and are standardized. The upper quartile and standard deviation of the null value set are recorded as the first anomaly threshold and the threshold gradient, respectively. The microprocessor region corresponding to the element in the null value set that is greater than the upper quartile is recorded as the first anomaly region. The difference between the gelatinization null value in the first anomaly region and the threshold gradient is the second anomaly threshold. If the set of gelatinization null values ​​corresponding to the eight neighbors of the first anomaly region is recorded as the gelatinization gradient set, and if more than half of the elements in the gelatinization gradient set are greater than the second anomaly threshold, then it is determined that a gelatinization anomaly has occurred in the first anomaly region; otherwise, no gelatinization anomaly has occurred.

5. The method for processing feed pellets using fermented silkworm excrement as raw material according to claim 4, characterized in that, In step S500, the method for identifying gelatinization abnormalities in feed based on gelatinization null values ​​further includes: screening the feed particles corresponding to the microprocessing area where gelatinization abnormalities occur, and then performing high-temperature steam gelatinization treatment.

6. A feed pellet processing system using fermented silkworm excrement as raw material, characterized in that, The feed pellet processing system using fermented silkworm excrement as raw material includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the feed pellet processing method using fermented silkworm excrement as raw material according to any one of claims 1-5. The feed pellet processing system using fermented silkworm excrement as raw material operates on a desktop computer, a laptop computer, a handheld computer, or a computing device in a cloud data center.

Images