Deer product production monitoring system based on machine vision

Abstract

The invention relates to the technical field of image processing, in particular to a deer product production monitoring system based on machine vision, which comprises an image acquisition module; the monitoring setting module is used for determining the monitoring interval duration according to the ossification pixel value proportion so as to obtain a plurality of monitoring points; the primary judgment module is used for primarily judging whether the production process is qualified or not according to the abnormal coefficient of the monitoring point; the secondary judgment module is used for determining, according to the sub-region anomaly difference degree, adjustment on the anomaly coefficient or key region analysis according to the difference comparison degree to obtain an adjustment anomaly coefficient, and secondarily judging whether the production process is qualified or not based on the adjustment anomaly coefficient; and the abnormity adjusting module is used for adjusting the heating rate or the heating temperature according to the abnormity comparison value and adjusting the stirring rate according to the proportion of the key area under the condition of primary adjustment instability. The method can guarantee the quality stability of pilose antler.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a machine vision-based deer product production monitoring system. Background Technology

[0002] In the deer product industry, deer antler slices are one of the core categories due to their extremely high medicinal and tonic effects. Since deer antler slices need to be boiled during processing, the heating process in the boiling stage plays a decisive role in the quality. However, improper heating often leads to quality risks such as loss of active ingredients and deterioration in appearance and color. Traditional production monitoring methods have problems such as subjective judgment and strong reliance on human intervention, resulting in poor quality stability and low pass rate of deer antler slices. Therefore, how to monitor the heating process to improve the quality of deer antler is a problem that urgently needs to be solved by those skilled in the art.

[0003] Chinese Patent Publication No. CN104127440A discloses a processing method for deer antler, including: 1) decocting red dates, filtering, and obtaining red date water; 2) soaking deer antler in the red date water, boiling, then removing the deer antler and draining it; 3) drying the drained deer antler. However, the above method has the following problems: it lacks real-time monitoring of the quality of deer antler during the heating process, cannot actively avoid quality risks caused by improper heating, and cannot guarantee the stability of deer antler quality. Summary of the Invention

[0004] To address this, the present invention provides a machine vision-based deer product production monitoring system to overcome the problems in the prior art, such as the lack of real-time monitoring of deer antler quality during the heating process, the inability to proactively avoid quality risks caused by improper heating, and the inability to guarantee the stability of deer antler quality.

[0005] To achieve the above objectives, the present invention provides a machine vision-based deer product production monitoring system, comprising: Image acquisition module, used to acquire images of deer products and infrared images; The monitoring setting module is connected to the image acquisition module and is used to determine the pixel category based on the pixel value and determine the monitoring interval duration based on the proportion of ossified pixel values ​​in the deer product image at the initial moment to obtain a number of monitoring points. The initial judgment module is connected to the image acquisition module and the monitoring setting module respectively. It is used to determine the abnormality coefficient of the monitoring point based on the difference in thermal expansion of ossification and the color decay rate of non-ossification, and to make an initial judgment on whether the production process is qualified based on the abnormality coefficient of the monitoring point. The secondary judgment module, which is connected to the primary judgment module, is used to determine whether the production process is qualified in the primary judgment when the production process is qualified in the primary judgment. Based on the sub-regional abnormality difference degree, the abnormality coefficient is adjusted according to the difference comparison degree or key area analysis is performed to obtain the adjusted abnormality coefficient, and the production process is qualified in the secondary judgment based on the adjusted abnormality coefficient. An abnormality adjustment module, which is connected to the initial judgment module and the secondary judgment module respectively, is used to adjust the heating rate or heating temperature according to the abnormality comparison value under abnormal conditions, and to adjust the stirring rate according to the proportion of the key area under the condition of first-level adjustment instability.

[0006] Furthermore, the monitoring setting module determines the pixel category based on the pixel value, including: For a single pixel in the antler slice area of ​​a deer product image, If the pixel value of the pixel is within the first preset pixel value range, then the pixel category of the pixel is skeletalized pixel. If the pixel value of a pixel is within the second preset pixel value range, then the pixel category of the pixel is a non-ossified pixel.

[0007] Furthermore, the monitoring setting module determines the monitoring interval duration based on the proportion of ossified pixel values ​​in the deer product image at the initial moment; The monitoring interval duration is negatively correlated with the proportion of ossified pixel values.

[0008] Furthermore, the initial judgment module determines whether the production process is qualified based on the anomaly coefficient of the monitoring points, including: For a single monitoring point, If the abnormality coefficient of the monitoring point is greater than or equal to the preset abnormality coefficient, the production process is initially determined to be unqualified. If the anomaly coefficient of the monitoring point is less than the preset anomaly coefficient, the production process is initially judged to be qualified. The abnormality coefficient is positively correlated with the difference in thermal expansion of ossification and the color decay rate of non-ossification.

[0009] Furthermore, if the sub-regional anomaly difference degree of the secondary judgment module is greater than or equal to the preset sub-regional anomaly difference degree, then the anomaly coefficient is increased and adjusted according to the difference comparison degree. The increase in the anomaly coefficient is positively correlated with the degree of difference comparison.

[0010] Furthermore, if the sub-region anomaly difference degree of the secondary judgment module is less than the preset sub-region anomaly difference degree, then key region analysis is performed. In the key area analysis, the key area is determined based on the bubble uniformity, bubble anomaly or pixel change value, and the sub-anomaly reference value is determined based on the infrared deviation value and the area characterization value of the key area. The anomaly reference value is determined based on the average value of the sub-anomaly reference values ​​corresponding to each key area, and the anomaly coefficient is increased based on the anomaly reference value. The increase in the abnormal coefficient is positively correlated with the abnormal reference value.

[0011] Furthermore, the secondary determination module determines key regions based on bubble uniformity, bubble anomaly, or pixel change values, including: If the bubble uniformity is less than the preset bubble uniformity, then the key area is determined based on the bubble anomaly. If the bubble uniformity is greater than or equal to the preset bubble uniformity, the key area is determined based on the pixel change value.

[0012] Furthermore, the secondary determination module determines the sub-anomaly reference value based on the infrared deviation value of the key area and the area characterization value; The sub-anomaly reference value is positively correlated with both the infrared deviation value and the regional characterization value.

[0013] Furthermore, the anomaly adjustment module reduces the heating rate or heating temperature based on the anomaly comparison value; The decrease in the heating rate is positively correlated with the anomaly comparison value, and the decrease in the heating temperature is positively correlated with the anomaly comparison value.

[0014] Furthermore, under the condition of a single adjustment instability, the abnormal adjustment module increases the stirring rate according to the proportion of the key area; The increase in the stirring rate is positively correlated with the proportion of the critical area.

[0015] Compared with the prior art, the beneficial effects of the present invention are that, in the technical solution of the present invention, the proportion of ossified pixel values ​​effectively reflects the differences in the distribution of components in deer antler slices, and then the monitoring interval is determined according to the proportion of ossified pixel values. This is conducive to achieving accurate matching between component risk and monitoring frequency, avoiding the problem of poor monitoring targeting in traditional fixed-interval monitoring, and improving the accuracy and efficiency of monitoring, thereby enhancing the stability of deer antler slice production quality.

[0016] Furthermore, in this invention, the difference in thermal expansion due to ossification and the rate of color decay due to non-ossification effectively reflect the quality change state of deer antler slices during the heating process. Then, based on the difference in thermal expansion due to ossification and the rate of color decay due to non-ossification, an anomaly coefficient is determined. The anomaly coefficient effectively reflects the comprehensive risk level of the heating process deviating from the historical qualified benchmark. Then, based on the anomaly coefficient, it is determined whether the production process is qualified, which helps to reduce human judgment error and improve the stability and consistency of deer antler slice production quality.

[0017] Furthermore, in this invention, the sub-regional anomaly difference degree effectively reflects the color uniformity of different sub-regions during the heating process of deer antler slices. Then, based on the sub-regional anomaly difference degree, the anomaly coefficient is adaptively increased or key region analysis is performed according to the difference comparison degree. This makes the determination of the anomaly coefficient more in line with the actual application scenario, avoids the risk of misjudgment caused by a single judgment standard, and thus ensures the stability of the production quality of deer antler slices.

[0018] Furthermore, under abnormal conditions, this invention adjusts the heating rate or heating temperature based on the abnormality comparison value, which helps to curb irreversible damage to the quality of deer antler slices caused by abnormal temperature. By precisely matching the severity of the abnormality to adjust the temperature parameters, it avoids degradation of active ingredients and abnormal color due to excessively high temperature, or aggravation of local temperature difference due to excessively rapid heating. It blocks the expansion of abnormality from the core variable level. Under conditions of single-adjustment instability, the stirring rate is adjusted according to the proportion of key areas, which helps to specifically solve the derivative problem of macroscopic temperature being qualified but microscopic uneven heating, and can avoid local quality abnormalities caused by uneven heating. Attached Figure Description

[0019] Figure 1 This is a module connection diagram of the deer product production monitoring system based on machine vision according to the present invention; Figure 2 This is a flowchart illustrating how the present invention determines the pixel category based on pixel values; Figure 3 This is a flowchart of the present invention for initially determining whether the production process is qualified based on the anomaly coefficient of the monitoring points; Figure 4 This is a flowchart illustrating how the present invention determines the adjustment of the anomaly coefficient or performs key region analysis based on the difference comparison degree according to the anomaly degree of sub-regions. Detailed Implementation

[0020] 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.

[0021] 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.

[0022] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate the direction or positional relationship, are based on the direction or positional relationship shown in the 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.

[0023] Please see Figures 1 to 4 As shown, the present invention provides a machine vision-based deer product production monitoring system, comprising: Image acquisition module, used to acquire images of deer products and infrared images; The monitoring setting module is connected to the image acquisition module and is used to determine the pixel category based on the pixel value and determine the monitoring interval duration based on the proportion of ossified pixel values ​​in the deer product image at the initial moment to obtain a number of monitoring points. The initial judgment module is connected to the image acquisition module and the monitoring setting module respectively. It is used to determine the abnormality coefficient of the monitoring point based on the difference in thermal expansion of ossification and the color decay rate of non-ossification, and to make an initial judgment on whether the production process is qualified based on the abnormality coefficient of the monitoring point. The secondary judgment module, which is connected to the primary judgment module, is used to determine whether the production process is qualified in the primary judgment when the production process is qualified in the primary judgment. Based on the sub-regional abnormality difference degree, the abnormality coefficient is adjusted according to the difference comparison degree or key area analysis is performed to obtain the adjusted abnormality coefficient, and the production process is qualified in the secondary judgment based on the adjusted abnormality coefficient. An abnormality adjustment module, which is connected to the initial judgment module and the secondary judgment module respectively, is used to adjust the heating rate or heating temperature according to the abnormality comparison value under abnormal conditions, and to adjust the stirring rate according to the proportion of the key area under the condition of first-level adjustment instability.

[0024] The application scenario of this invention is the monitoring of the heating process of deer antler slices. This invention includes several historical records, each recording at least one instance of monitoring during the heating process, including the ossification thermal expansion coefficient, non-ossification color decay rate, sub-regional abnormality, and abnormal reference values. Each historical record also has a corresponding pass / fail marker, indicating whether the monitoring process during the deer antler slice heating meets user requirements. The pass / fail marker can be manually recorded. It is understood that users can determine whether the monitoring process during the deer antler slice heating meets their requirements based on self-defined indicators. These self-defined indicators can be, but are not limited to, the slice integrity rate, which will not be elaborated here. Specifically, after the deer antler slices are heated, 100 slices are randomly selected. The slice integrity rate is calculated as: (Number of deer antler slices without breakage or edge defects) / 100. The heating process of deer antler slices includes: laying the deer antler slices in a single layer in a pot of purified water at room temperature, slowly heating them to 80℃ at a heating rate of 2℃ / min, and maintaining the temperature at 80℃ for 30 minutes, stirring with a stirrer. The stirrer blades are installed 12cm above the bottom of the pot to avoid directly hitting the deer antler slices. The stirring speed is 30rpm. There are no restrictions on the specific model of the stirrer, and users can choose according to their actual needs. The deer product image is a denoised and grayscale image taken by a camera 1.5m directly above the cooking pot. The infrared thermal imager and the camera are fixed on the same bracket to ensure that the shooting angles of the two are completely consistent. The infrared image is enlarged to the same pixel size as the deer product image to obtain a preprocessed infrared image. The methods that can be used to enlarge the infrared image to the same pixel size as the deer product image include, but are not limited to, MATLAB and OpenCV. Users can choose according to their actual needs, and there are no restrictions on the specific methods. Abnormal conditions include both initial determination that the production process is unqualified and secondary determination that the production process is unqualified.

[0025] Specifically, the monitoring setting module determines the pixel category based on the pixel value, including: For a single pixel in the antler slice area of ​​a deer product image, If the pixel value of the pixel is within the first preset pixel value range, then the pixel category of the pixel is skeletalized pixel. If the pixel value of a pixel is within the second preset pixel value range, then the pixel category of the pixel is a non-ossified pixel.

[0026] Among them, the antler slice region is the pixel region in the deer product image that contains only the antler slice itself. The region where the antler slice is located in the deer product image is identified by OpenCV. It should be noted that the first preset pixel value corresponding to the ossified pixels in the deer product images corresponding to different monitoring points is different, and the second preset pixel value corresponding to the non-ossified pixels in the deer product images corresponding to different monitoring points is also different. For the deer product image corresponding to a single monitoring point, any pixel value in the first preset pixel value range is less than or equal to the first preset pixel value, and any pixel value in the second preset pixel value range is greater than or equal to the second preset pixel value. The first preset pixel value corresponding to a single monitoring point is less than the second preset pixel value; The user can determine the first and second preset pixel values ​​corresponding to a single monitoring point based on the actual application scenario. The stricter the user's determination of the pixel category, the smaller the first preset pixel value and the larger the second preset pixel value. A method is provided whereby the maximum value among the ossified pixels in the deer product image corresponding to the monitoring point in the historical record (with the same time elapsed since the initial time) that meets the user's needs is recorded as the first preset pixel value, and the minimum value among the non-ossified pixels in the deer product image corresponding to the monitoring point in the historical record (with the same time elapsed since the initial time) that meets the user's needs is recorded as the second preset pixel value.

[0027] Specifically, the monitoring setting module determines the monitoring interval duration based on the proportion of ossified pixel values ​​in the deer product image at the initial moment; The monitoring interval duration is negatively correlated with the proportion of ossified pixel values.

[0028] The initial moment is the moment after the deer antler slices are laid flat in a single layer and placed in a pot of room temperature purified water, but before heating begins; Ossified pixel value ratio = number of ossified pixels in the antler slice region of the deer product image at the initial moment / total number of ossified and non-ossified pixels in the antler slice region of the deer product image at the initial moment; Monitoring interval duration = t - t0 × percentage of boned pixel values, where t is 10 min and t0 is 5 min; Starting from the initial time, an interval point is set at each monitoring interval until heating is completed, and each set interval point is recorded as a monitoring point. Specifically, the initial judgment module determines whether the production process is qualified based on the anomaly coefficient of the monitoring points, including: For a single monitoring point, If the abnormality coefficient of the monitoring point is greater than or equal to the preset abnormality coefficient, the production process is initially determined to be unqualified. If the anomaly coefficient of the monitoring point is less than the preset anomaly coefficient, the production process is initially judged to be qualified. The abnormality coefficient is positively correlated with the difference in thermal expansion of ossification and the color decay rate of non-ossification.

[0029] Among them, the difference in ossification thermal expansion corresponding to a single monitoring point = |ossification thermal expansion coefficient - the average value of the ossification thermal expansion coefficients corresponding to each monitoring point in the historical records that meet the user's needs and are qualified in the second judgment |. The coefficient of thermal expansion of ossification corresponding to a single monitoring point = (the number of ossified pixels in the antler slice region of the deer product image corresponding to the monitoring point - the number of ossified pixels in the antler slice region of the deer product image corresponding to the initial time) / the time length from the monitoring point to the initial time; The non-ossified color decay rate corresponding to a single monitoring point = (a1-a2) / the time length from the monitoring point to the initial time, where the unit of the time length from the monitoring point to the initial time is min. The average pixel value of each non-ossified pixel in the antler slice area of ​​the deer product image corresponding to the initial time is recorded as a1, and the average pixel value of each non-ossified pixel in the antler slice area of ​​the deer product image corresponding to the monitoring point is recorded as a2. The anomaly coefficient corresponding to a single monitoring point is y0, where y0 = (the difference in thermal expansion of ossification corresponding to the monitoring point / the preset difference in thermal expansion of ossification) × the first weighting coefficient + (the non-ossification color decay rate corresponding to the monitoring point / the preset non-ossification color decay rate) × the second weighting coefficient. The first weighting coefficient is 0.4, and the second weighting coefficient is 0.6. The values ​​of the preset osteoclast thermal expansion difference and the preset non-ossified color decay rate can be determined by the user according to the actual application scenario. The greater the user's requirement for improving the accuracy of the heating process quality stability, the smaller the values ​​of the preset osteoclast thermal expansion difference and the preset non-ossified color decay rate. A method for determining the values ​​of the preset osteoclast thermal expansion difference and the preset non-ossified color decay rate is provided. The average value of the osteoclast thermal expansion difference and the average value of the non-ossified color decay rate corresponding to the historical records that can meet the user's needs are detected and recorded as the preset osteoclast thermal expansion difference and the preset non-ossified color decay rate, respectively. The user can determine the value of the preset anomaly coefficient according to the actual application scenario. The greater the user's need to improve the accuracy of anomaly monitoring, the smaller the value of the preset anomaly coefficient. One preset anomaly coefficient value is provided, which is 0.5. Understandably, the ossified areas form the supporting structure of the antler slices. Large differences in thermal expansion during ossification indicate that these areas may over-expand, crack, or deform during heating. The non-ossified areas of the antler slices are rich in active ingredients such as proteins and peptides, and their color is strongly correlated with the retention of these components: a high rate of color decay in non-ossified areas often indicates that active ingredients have degraded, oxidized, or been lost during heating. Therefore, the anomaly coefficient determined by the difference in thermal expansion during ossification and the rate of color decay in non-ossified areas can reflect the risk of the antler slices deviating from historical acceptable standards during the heating process.

[0030] Specifically, if the sub-regional anomaly difference degree of the secondary judgment module is greater than or equal to the preset sub-regional anomaly difference degree, then the anomaly coefficient is increased and adjusted according to the difference comparison degree. The increase in the anomaly coefficient is positively correlated with the degree of difference comparison.

[0031] Among them, for the deer product image corresponding to a single monitoring point, it is divided into several rectangular regions with the same area and adjacent to each other. The rectangular regions that overlap with the deer antler slice region are recorded as sub-regions. The sub-region abnormality degree is the standard deviation of the pixel mean corresponding to each sub-region, and the average pixel value corresponding to each pixel in the deer antler slice region in a single sub-region. The value of the preset sub-region anomaly difference degree can be determined by the user according to the actual application scenario. The greater the user's demand for improving monitoring accuracy, the smaller the value of the preset sub-region anomaly difference degree. A method for determining the value of the preset sub-region anomaly difference degree is provided, which is the average value of the sub-region anomaly difference degree corresponding to each monitoring point in the historical records that can meet the user's needs and determine the production process as qualified in the second determination. Difference comparison score = Sub-region abnormal difference score - Preset sub-region abnormal difference score; The increase in the anomaly coefficient = y0 × (sub-region anomaly difference - preset sub-region anomaly difference) / preset sub-region anomaly difference; Adjustment abnormality coefficient = increase in abnormality coefficient + y0; Specifically, if the sub-region anomaly difference degree of the secondary judgment module is less than the preset sub-region anomaly difference degree, then key region analysis is performed. In the key area analysis, the key area is determined based on the bubble uniformity, bubble anomaly or pixel change value, and the sub-anomaly reference value is determined based on the infrared deviation value and the area characterization value of the key area. The anomaly reference value is determined based on the average value of the sub-anomaly reference values ​​corresponding to each key area, and the anomaly coefficient is increased based on the anomaly reference value. The increase in the abnormal coefficient is positively correlated with the abnormal reference value.

[0032] Wherein, the bubble uniformity corresponding to a single monitoring point = 1 / (standard deviation of bubble reference values ​​corresponding to each sub-region in the deer product image corresponding to the monitoring point + 1). The bubbles in the cooking pot area and each sub-region in the deer product image are identified by OpenCV. This is a common technique used by those skilled in the art, and will not be elaborated on in detail. The cooking pot area is the pixel range corresponding to the cooking pot in the deer product image. The bubble reference value for a single sub-region = the number of bubbles in that sub-region / the area where that sub-region overlaps with the cooking pot area; the number of bubbles in a single sub-region is the number of bubbles in that sub-region. Bubble anomaly degree for a single sub-region = |bubble reference value for this sub-region - average value of bubble reference values ​​for all sub-regions| / (average value of bubble reference values ​​for all sub-regions + 1); The pixel change value corresponding to a single sub-region = |Standard deviation of pixel values ​​corresponding to each pixel in the antler slice region of this sub-region -Standard deviation of pixel values ​​corresponding to each pixel in the antler slice region of the deer product image|; The increase in the anomaly coefficient = y0 × (anomaly reference value / average of the anomaly reference values ​​corresponding to historical records that can meet user needs); Adjustment abnormality coefficient = increase in abnormality coefficient + y0; If the adjustment abnormality coefficient is greater than or equal to the preset abnormality coefficient, the production process will be deemed unqualified in the second judgment. If the adjustment abnormality coefficient is less than the preset abnormality coefficient, the production process is deemed qualified in the second judgment. The anomaly reference value for a single monitoring point is the average of the sub-anomaly reference values ​​corresponding to each key area in the deer product image corresponding to that monitoring point.

[0033] It is understandable that the sub-region abnormality difference degree can effectively reflect the color difference of deer antler slices in different sub-regions. When the sub-region abnormality difference degree is greater than or equal to the preset sub-region abnormality difference degree, it indicates that the color difference of deer antler slices in different sub-regions is large. Therefore, the abnormality coefficient is increased and adjusted according to the difference comparison degree. When the abnormal difference degree of a sub-region is less than the preset abnormal difference degree of a sub-region, it indicates that the color difference of the deer antler slices in different sub-regions is small. By conducting key area analysis, even if the color has not yet shown significant differences, potential risks can be identified in advance through abnormal bubbles, temperature deviations, etc.

[0034] Specifically, the secondary determination module determines key areas based on bubble uniformity, bubble anomaly, or pixel change values, including: If the bubble uniformity is less than the preset bubble uniformity, then the key area is determined based on the bubble anomaly. If the bubble uniformity is greater than or equal to the preset bubble uniformity, the key area is determined based on the pixel change value.

[0035] The user can determine the value of the preset bubble uniformity according to the actual application scenario. The larger the value of the preset bubble uniformity, the greater the user's need to determine the key area based on the bubble anomaly. A method for determining the value of the preset bubble uniformity is provided, which detects the historical records of the user determining the key area based on the bubble anomaly, and records the average value of the bubble uniformity corresponding to the historical records that can meet the user's needs as the preset bubble uniformity. When determining key regions based on bubble anomaly, sub-regions with bubble anomaly greater than the preset bubble anomaly are recorded as key regions. When determining key regions based on pixel change values, sub-regions with pixel change values ​​greater than preset pixel change values ​​are recorded as key regions. The preset values ​​of bubble anomaly and preset pixel change value can be determined by the user according to the actual application scenario. The greater the user's need to improve monitoring accuracy, the smaller the preset values ​​of bubble anomaly and preset pixel change value should be. A method for determining the preset values ​​of bubble anomaly and preset pixel change value is provided, which is the average value of the bubble anomaly corresponding to each key area in the historical records that can meet the user's needs and determine the key area based on bubble anomaly, and the average value of the pixel change value corresponding to each key area in the historical records that can meet the user's needs and determine the key area based on pixel change value, is recorded as the preset pixel change value.

[0036] Specifically, the secondary determination module determines the sub-anomaly reference value based on the infrared deviation value of the key area and the area characterization value; The sub-anomaly reference value is positively correlated with both the infrared deviation value and the regional characterization value.

[0037] For deer product images and infrared images of a single monitoring point, the infrared reference value of a single sub-region at the monitoring point is the average value of the pixel values ​​of each pixel point in the preprocessed infrared image that has the same coordinates and size as the sub-region. The infrared deviation value of a single key area at a single monitoring point = |Infrared reference value of the key area at the monitoring point - Average value of infrared reference values ​​of each sub-area at the monitoring point|; Sub-anomaly reference value = Infrared deviation value / Preset infrared deviation value + Area characterization value / Preset area characterization value; The region characterization value corresponding to a single key region = bubble anomaly degree / preset bubble anomaly degree + pixel change value / preset pixel anomaly degree; The user can determine the preset infrared deviation value and preset area characterization value according to the actual application scenario. The greater the user's demand for improving the monitoring accuracy of the heating process, the smaller the preset infrared deviation value and preset area characterization value should be. One preset infrared deviation value and preset area characterization value is provided. The average value of the infrared deviation value corresponding to the historical records that can meet the user's needs is recorded as the preset infrared deviation value, and the preset area characterization value is 1.4. Specifically, the anomaly adjustment module reduces the heating rate or heating temperature based on the anomaly comparison value. The decrease in the heating rate is positively correlated with the anomaly comparison value, and the decrease in the heating temperature is positively correlated with the anomaly comparison value.

[0038] It should be noted that for a single monitoring point, if the heating temperature reaches 80℃ from the initial moment to that monitoring point, the heating temperature will be reduced; if the heating temperature does not reach 80℃, the heating rate will be reduced. The heating temperature is the water temperature at any point in the pot, measured by a temperature sensor. Anomaly comparison value = y - preset anomaly coefficient; If the production process is initially determined to be unqualified, then y is the anomaly coefficient; If the production process is deemed unqualified in the second determination, then y is the adjustment anomaly coefficient; The decrease in heating rate = anomaly comparison value / preset anomaly coefficient × heating rate threshold, where the heating rate threshold is 1℃ / min; The decrease in heating temperature = anomaly comparison value / preset anomaly coefficient × heating temperature threshold, where the heating temperature threshold is 10℃; Understandably, if the heating rate is too fast, it will exacerbate the temperature gradient between different sub-regions, further amplifying the differences in color and texture. If the heating temperature is too high, it may quickly cause the local color of the antler slices to darken and the internal moisture to evaporate rapidly, forming a large number of bubbles. In this case, the heating rate or heating temperature should be adjusted first. By adjusting the heating rate or heating temperature, the continuous damage to the quality of the antler slices caused by temperature imbalance can be directly alleviated.

[0039] Specifically, under the condition of a single adjustment instability, the abnormal adjustment module increases the stirring rate according to the proportion of the key area; The increase in the stirring rate is positively correlated with the proportion of the critical area.

[0040] Among them, the first adjustment instability condition is that after 2 minutes of adjusting the heating rate to reduce the abnormal comparison value, the proportion of the critical area is greater than the preset proportion of the critical area. The time 2 minutes after the heating rate was reduced based on the abnormal comparison value was recorded as the reference time. Key region percentage = Number of key regions in the deer product image taken at the reference time / Number of sub-regions in the deer product image taken at the reference time; The user can determine the value of the preset critical area percentage based on the actual application scenario. The greater the user's need for improved accuracy in production stability, the smaller the value of the preset critical area percentage should be. One preset critical area percentage is provided, which is 40%. The increase in stirring rate = percentage of critical region / threshold of critical region percentage × stirring rate threshold; the threshold of critical region percentage is 1, and the threshold of ultrasonic oscillation frequency is 20 rpm; It should be noted that if the proportion of the critical area is still greater than the preset critical area proportion after adjusting the stirring rate according to the proportion of the critical area, heating should be stopped to avoid continuous abnormal processing that could damage the quality of the deer products. Understandably, by adjusting the heating rate or heating temperature based on the abnormal comparison value, the heating rate and heating temperature of the cooking pot have been controlled within a reasonable range. The instability condition after one adjustment indicates that the deer antler slices in the cooking pot are not heated evenly. The more sub-regions with quality abnormalities in the deer product image, the more serious the problem of uneven heating of the material is. Stronger stirring intervention is needed to break the local temperature difference to ensure the stability of the deer product production quality.

[0041] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A machine vision-based deer product production monitoring system, characterized by, The method comprises the following steps: an image acquisition module is configured to acquire images of deer products and infrared images; a monitoring setting module is connected to the image acquisition module and configured to determine pixel point categories according to pixel values and determine a monitoring interval duration according to a proportion of ossification pixel values in the deer product images at an initial time to obtain a plurality of monitoring points; a first judgment module is connected to the image acquisition module and the monitoring setting module and configured to determine an abnormality coefficient of the monitoring points according to a heat expansion difference of ossification and a color and luster attenuation rate of non-ossification and initially judge whether the production process is qualified according to the abnormality coefficient of the monitoring points; a second judgment module is connected to the first judgment module and configured to, when the production process is initially judged to be qualified, determine a difference comparison degree according to a sub-region abnormality difference, adjust the abnormality coefficient according to the difference comparison degree or perform key region analysis to obtain an adjusted abnormality coefficient, and secondarily judge whether the production process is qualified based on the adjusted abnormality coefficient; an abnormality adjustment module is connected to the first judgment module and the second judgment module and configured to, under abnormal conditions, adjust a temperature rising rate or a heating temperature according to an abnormality comparison value, and, under a condition of instability after a first adjustment, adjust a stirring rate according to a key region proportion.

2. The machine vision-based deer product production monitoring system of claim 1, wherein, The monitoring setting module determines pixel point categories according to pixel values, which comprises: for a single pixel point in a region of deer horn slices in the deer product images, if the pixel value of the pixel point is in a first preset pixel value range, the pixel point category of the pixel point is an ossification pixel point; if the pixel value of the pixel point is in a second preset pixel value range, the pixel point category of the pixel point is a non-ossification pixel point.

3. The machine vision-based deer product production monitoring system of claim 2, wherein, The monitoring setting module determines a monitoring interval duration according to a proportion of ossification pixel values in the deer product images at an initial time. The monitoring interval duration and the proportion of ossification pixel values are in a negative correlation relationship.

4. The machine vision-based deer product production monitoring system of claim 3, wherein, The first judgment module determines whether the production process is qualified according to the abnormality coefficient of the monitoring points, which comprises: for a single monitoring point, if the abnormality coefficient of the monitoring point is greater than or equal to a preset abnormality coefficient, the production process is initially judged to be unqualified; if the abnormality coefficient of the monitoring point is less than the preset abnormality coefficient, the production process is initially judged to be qualified; the abnormality coefficient is in a positive correlation relationship with the heat expansion difference of ossification and the color and luster attenuation rate of non-ossification.

5. The machine vision-based deer product production monitoring system of claim 1, wherein, The second judgment module, in response to a sub-region abnormality difference greater than or equal to a preset sub-region abnormality difference, adjusts the abnormality coefficient according to a difference comparison degree; the increase value of the abnormality coefficient and the difference comparison degree are in a positive correlation relationship.

6. The machine vision-based deer product production monitoring system of claim 1, wherein, The second judgment module, in response to a sub-region abnormality difference less than a preset sub-region abnormality difference, performs key region analysis; in the key region analysis, a key region is determined according to a bubble uniformity or a pixel change value, a sub-abnormality reference value is determined according to an infrared deviation value and a region representation value of the key region, an abnormality reference value is determined based on an average value of the sub-abnormality reference values corresponding to each key region, and the abnormality coefficient is adjusted according to the abnormality reference value; the increase value of the abnormality coefficient and the abnormality reference value are in a positive correlation relationship.

7. The machine vision-based deer product production monitoring system of claim 6, wherein, The secondary determination module determines the key region according to the bubble uniformity, the bubble abnormality, or the pixel change value, comprising: If the bubble uniformity is less than a preset bubble uniformity, the key region is determined according to the bubble abnormality; If the bubble uniformity is greater than or equal to the preset bubble uniformity, the key region is determined according to the pixel change value.

8. The machine vision-based deer product production monitoring system of claim 6, wherein, The secondary determination module determines the sub-abnormal reference value according to the infrared deviation value and the region representation value of the key region; The sub-abnormal reference value is positively correlated with the infrared deviation value and the region representation value.

9. The machine vision-based deer product production monitoring system of claim 1, wherein, The abnormality adjustment module adjusts the heating rate or the heating temperature according to the abnormality comparison value; The decrease value of the heating rate is positively correlated with the abnormality comparison value, and the decrease value of the heating temperature is positively correlated with the abnormality comparison value.

10. The machine vision-based deer product production monitoring system of claim 9, wherein, The abnormality adjustment module adjusts the stirring rate according to the key region proportion under the condition of instability of the primary adjustment; The increase value of the stirring rate is positively correlated with the key region proportion.

Images