Classical swine fever virus detection system based on cloud analysis

Abstract

The invention discloses a hog cholera virus detection system based on cloud analysis, which comprises an image acquisition unit, a data processing cloud, a disease monitoring unit and an early warning management unit, and is characterized in that a live pig limb image is acquired through the image acquisition unit, and an image analysis model is constructed through the data processing cloud to analyze the live pig limb image; the hog cholera risk suspected degree is comprehensively evaluated from two perspectives of vascular risk and tissue edema, so that auxiliary diagnosis is carried out on hog cholera virus detection; the disease monitoring unit is used for monitoring the hog cholera virus for a long time, the deterioration or recovery condition of the disease condition of the sick pig is analyzed, a sick pig health management scheme is generated, and the targeted treatment effect of the sick pig is improved; and the early warning management unit globally analyzes the virus risk and generates a live pig breeding management scheme, so that corresponding overall management and control are performed for the hog cholera virus infection degree of the live pig breeding base, the early warning management efficiency of the live pig breeding base is improved, and the live pig breeding loss and the management cost are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of virus detection, and in particular to a classical swine fever virus detection system based on cloud analysis. Background Art

[0002] Classical swine fever virus is the pathogen of classical swine fever. Classical swine fever is an acute, febrile, highly contagious disease that is extremely harmful to pigs and will cause significant losses to the pig farming industry. After being infected with the classical swine fever virus, pigs will show obvious fluid accumulation or local inflammation in their limbs and other parts, and their blood vessels will become red, hot and swollen, and may be locally bleeding, infarcted, or necrotic, indicating a serious infection or complication. Therefore, pigs with the above symptoms are judged to be suspected of being infected with the classical swine fever virus. The existing method of detecting the swine fever virus is mainly through biomedical testing technology, collecting samples from suspected sick pigs and then conducting experimental testing. There are technical defects such as high virus testing costs, large demand for manpower sampling, and easy spread of the virus. Due to the sudden outbreak of swine fever and insufficient manpower in pig breeding bases, the efficiency of swine fever virus diagnosis is low. The swine fever virus is highly infectious and may miss the best time for control. In addition, traditional biological sample collection is mainly saliva or blood samples. The collection process is also prone to virus spread, leading to the continued deterioration of the swine fever epidemic. In view of the above technical defects, a solution is now proposed. Summary of the invention

[0003] The purpose of the present invention is to solve the technical defects of the existing classical swine fever virus detection methods, such as high virus detection cost, large demand for manpower sampling, and easy virus spread and infection. By remotely collecting images of pig limbs and sending them to the cloud for data processing, the demand for manpower sampling is reduced, and an image analysis model is constructed to evaluate the suspicion of classical swine fever virus infection, thereby improving the work efficiency of virus detection. The accuracy of the image analysis model is compared and tested by sampling biological detection, thereby continuously optimizing the image analysis model, improving the accuracy of virus detection, significantly reducing the trend of virus infection spread, and effectively curbing the continued deterioration of the classical swine fever epidemic.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions: A swine fever virus detection system based on cloud analysis, comprising an image acquisition unit, a data processing cloud, a disease monitoring unit and an early warning management unit, wherein the image acquisition unit, the data processing cloud, the disease monitoring unit and the early warning management unit are communicatively connected; The image acquisition unit is used to collect images of pig limbs: the image acquisition period Tp is set, and the wearable device is bound to the pig limbs, and an infrared image acquisition device is set in the wearable device. The infrared image acquisition device is used to regularly acquire images of pig limbs, and then the images of pig limbs are transmitted to the data processing cloud through remote signals; The data processing cloud is used to build image analysis models to analyze pig limb images: the data processing cloud includes a vascular risk analysis module, a tissue edema analysis module, a swine fever risk assessment module, and a comparison optimization module; The blood flow of limb blood vessels is analyzed through the vascular risk analysis module to assess the degree of vascular risk; the tissue edema analysis module is used to analyze the tissue edema of limb parts and assess the degree of tissue edema; the swine fever risk assessment module combines the vascular risk degree with the tissue edema degree to comprehensively assess the suspicion of swine fever and obtain the image detection results of swine fever virus, thereby providing auxiliary diagnosis for swine fever virus detection; the comparison and optimization module obtains the biological detection results of swine fever virus, and feeds them back to the image analysis model for comparison with the image detection results to determine the detection accuracy of the image analysis model, thereby continuously optimizing the image analysis model; The disease monitoring unit is used for long-term monitoring of the classical swine fever virus: by regularly monitoring the image detection results of the classical swine fever virus, analyzing the long-term development status of the classical swine fever virus, assessing the health risk level of sick pigs, generating a health management plan for sick pigs and providing corresponding risk warnings; The early warning management unit is used to conduct a global analysis of virus risks: by conducting an overall analysis of the virus risk level, evaluating the infection status of the swine fever virus in the pig breeding base, generating a pig breeding management plan and conducting corresponding early warning management.

[0005] Furthermore, the specific construction process of the image analysis model is as follows: The image analysis model includes a vascular analysis sub-model, a tissue analysis sub-model and a comprehensive analysis sub-model; Sa, establishes a vascular analysis sub-model through the vascular risk analysis module to analyze the blood flow of limb blood vessels and assess the degree of vascular risk; Sb, establishes a tissue analysis sub-model through the tissue edema analysis module to analyze the tissue edema of the limbs and evaluate the degree of tissue edema; Sc, a comprehensive analysis sub-model is established through the swine fever risk assessment module, which combines the vascular risk level with the tissue edema level, comprehensively assesses the suspicion of swine fever, and obtains the image detection results of swine fever virus, thereby assisting the diagnosis of swine fever virus detection; Sd, obtains the biological detection results of classical swine fever virus through the comparison optimization module, and feeds it back to the image analysis model for comparison with the image detection results to determine the detection accuracy of the image analysis model, thereby continuously optimizing the image analysis model.

[0006] Furthermore, the specific process of establishing the vascular analysis sub-model is as follows: Sa-1, input pig limb images to the vascular analysis sub-model; Extract N0 pixels of the pig limb image and collect grayscale, and mark the grayscale value of any pixel i as HDi; Set the grayscale interval Qh of the blood vessel, extract the pixel points whose grayscale values ​​HDi are in the grayscale interval Qh, and integrate them into a blood vessel feature point set, and mark the area where the blood vessel feature point set is located as a blood vessel area; The blood vessel area is divided into N1 characteristic regions and the regional state analysis is performed. Any characteristic region is marked as J, and the grayscale mean η1 of n0 pixels in the characteristic region J is calculated; Set the grayscale interval Qη of regional infarction. When the grayscale mean η1 of the feature region J is within the grayscale interval Qη, the feature region is marked as the infarction region; otherwise, the feature region is marked as the blood flow region. Sa-2, conduct targeted analysis on the infarct area and blood flow area respectively. The specific process is as follows: Sa-201, analysis of infarct area: by calculating the proportion of the total infarct area in the vascular area, the infarct proportion η2 was obtained; Sa-202, blood flow area analysis: extract the pig limb images corresponding to N2 image acquisition cycles Tp, and construct a dynamic curve S0 between the grayscale mean η1 of the blood flow area and the image acquisition cycle Tp, and obtain the blood pumping cycle Tb and blood pumping intensity Db through the dynamic curve S0; Sa-3, by combining the infarction proportion η2, pumping cycle Tb and pumping intensity Db, the vascular risk assessment coefficient Xxg is obtained.

[0007] Furthermore, the specific process of establishing the organizational analysis sub-model is as follows: Sb-1, extracting the pixel points whose grayscale values ​​HDi of the pig limb image are not located in the grayscale interval Qh, and marking the area where the pixel points are located as the tissue area; Sb-2, divide the tissue area into M1 characteristic regions and perform regional state analysis, mark any characteristic region as R, and calculate the grayscale mean φ1 of m0 pixels in the characteristic region R; Sb-201, set the grayscale interval Qs of regional edema, and if the grayscale mean φ1 of the extracted feature region R is in the grayscale interval Qs, then mark the feature region as an edema region; Sb-202, by cumulatively measuring the proportion of all edema areas in the tissue area, obtaining the edema proportion φ2, extracting the pig limb images corresponding to N2 image acquisition cycles Tp, and obtaining the edema proportion of the pig limb images corresponding to N2 image acquisition cycles Tp; Sb-203, obtain the edema proportion growth rate ψ through the edema proportion φ2 corresponding to two adjacent image acquisition cycles, and then obtain the average edema growth rate Ψr and the edema variation coefficient σr; Sb-3, by combining the average edema increase rate Ψr and the edema variation coefficient σr, the tissue risk assessment coefficient Xzh is obtained.

[0008] Furthermore, the specific process of establishing the comprehensive analysis sub-model is as follows: Sc-1, constructing a change curve s1 between the vascular risk assessment coefficient Xxg and the image acquisition period Tp, and a change curve s2 between the tissue risk assessment coefficient Xzh and the image acquisition period Tp; Synchronously extract M3 points of the change curve s1 and the change curve s2 to establish the risk parameter matrix V; Sc-2, by obtaining the overall risk index va and the risk fluctuation index vb from the risk parameter matrix V, and then obtaining the risk factor υ, the risk factors of the vascular risk assessment coefficient Xxg and the tissue risk assessment coefficient Xzh are marked as υ1 and υ2 respectively; Sc-3, by combining the risk factor υ1 of the vascular risk assessment coefficient Xxg and the risk factor υ2 of the tissue risk assessment coefficient Xzh, the suspected reference index Zw of swine fever is obtained; A determination interval Qw of the swine fever suspected reference index Zw is set. When the swine fever suspected reference index Zw is within the determination interval Qw, the image detection result is determined to be infected with swine fever virus; otherwise, the image detection result is determined to be not infected with swine fever virus.

[0009] Furthermore, the specific optimization process of the image analysis model is as follows: Sd-1, mark the total pig population as ZT, set the sampling ratio q%, obtain the pig sample volume YB, select YB pig samples by random sampling, perform biological tests on the pig samples, obtain the biological test results, set the biological test value Ja, and assign the biological test value Ja according to the biological test results; Sd-2, set the image detection value Pa, and assign the image detection value Pa according to the image detection result; Sd-3, compare the biological detection value Ja and the image detection value Pa to obtain the detection deviation coefficient JP; Sd-4, set the evaluation interval of the detection deviation coefficient JP, evaluate the detection accuracy of the image analysis model through interval comparison, and then set the correction amplitude of the image analysis model to optimize and adjust the image analysis model.

[0010] Furthermore, the specific process of long-term monitoring of swine fever virus is as follows: When both the image detection result and the biological detection result of the classical swine fever virus are infected with the classical swine fever virus, the live pig sample is marked as a sick pig, the number of sick pigs is marked as BZ, and the classical swine fever suspected reference index Zw corresponding to the image detection result of the sick pig is extracted; Set the virus monitoring period Td, construct the change curve sd between the suspected reference index Zw of swine fever and the virus monitoring period Td, extract n3 points of the change curve sd and their coordinates, calculate and obtain the slope value corresponding to the n3 points, and obtain the curve increase rate Ψs and curvature fluctuation coefficient σs of the change curve sd; When the slope value Kpo of any point is 0, the point is determined to be a marked point, and the curve segment sz between the marked point Yz and the end point of the change curve sd is extracted to obtain the drop value Lz and the change rate Kz of the curve segment sz; Then, the health risk index Zjk of sick pigs is comprehensively obtained, and the evaluation interval of the health risk index Zjk of sick pigs is set. The health risk degree of sick pigs is evaluated through interval comparison, and the corresponding sick pig health management plan is generated and risk warnings are issued.

[0011] Furthermore, the specific process of conducting a global analysis of virus risks is as follows: The estimated number of sick pigs in the total pig volume ZT is obtained by measuring the proportion of the number of sick pigs BZ detected in the pig sample volume YB. Combined with the health risk index Zjk of sick pigs, the virus risk level of all pigs in the pig breeding base is analyzed as a whole to obtain the swine fever virus infection index Zgr; Set the evaluation range of the classical swine fever virus infection index Zgr, evaluate the degree of classical swine fever virus infection in the pig breeding base through interval comparison, generate the corresponding pig breeding management plan and conduct early warning management.

[0012] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: The present invention remotely collects images of pig limbs and sends them to the cloud for data processing, thereby reducing the demand for manual sampling, and constructs an image analysis model to evaluate the suspected degree of swine fever virus infection, thereby improving the efficiency of virus detection. The accuracy of the image analysis model is compared and tested by sampling biological detection, thereby continuously optimizing the image analysis model, improving the accuracy of virus detection, significantly reducing the trend of virus infection spread, and effectively curbing the continued deterioration of the swine fever epidemic. The present invention collects images of pig limbs through an image acquisition unit, and constructs an image analysis model on a data processing cloud to analyze the images of pig limbs, comprehensively assesses the suspected risk of classical swine fever from the two perspectives of vascular risk and tissue edema, thereby performing auxiliary diagnosis for classical swine fever virus detection; and then performs long-term monitoring of classical swine fever virus through a disease monitoring unit, analyzes the deterioration or recovery of the condition of sick pigs, generates a health management plan for sick pigs, thereby performing corresponding health management according to the health risk degree of sick pigs, and improving the targeted treatment effect of sick pigs; The present invention conducts a global analysis of virus risks through an early warning management unit and generates a pig breeding management plan, thereby performing corresponding overall control on the infection level of classical swine fever virus in the pig breeding base, improving the early warning management efficiency of the pig breeding base, and reducing pig breeding losses and management costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 Shows a schematic diagram of the modules of the present invention; Figure 2 A schematic diagram of the process of the present invention is shown. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0015] Embodiment 1: like Figure 1-Figure 2 As shown, a swine fever virus detection system based on cloud analysis includes an image acquisition unit, a data processing cloud, a disease monitoring unit and an early warning management unit, wherein the image acquisition unit, the data processing cloud, the disease monitoring unit and the early warning management unit are communicatively connected; S1, the image acquisition unit acquires images of pig limbs: the image acquisition period Tp is set, the wearable device is bound to the pig limbs, an infrared image acquisition instrument is set in the wearable device, the infrared image acquisition instrument is used to regularly acquire images of pig limbs, and then the images of pig limbs are remotely transmitted to the data processing cloud; Wearable devices can be used to conduct targeted health monitoring of pigs in pig breeding bases. Wearable devices use infrared sensors for optical image monitoring. Wearable devices can monitor the body temperature, heart rate, blood pressure and blood vessel conditions of pigs for a long time, so as to comprehensively analyze the physical health status of pigs. In the case of swine fever virus infection, abnormalities of pigs can be discovered in time, and remote monitoring of sick pigs can be achieved to prevent the spread of infection caused by humans carrying the virus. S2, the data processing cloud constructs an image analysis model to analyze pig limb images: the data processing cloud includes a vascular risk analysis module, a tissue edema analysis module, a swine fever risk assessment module, and a comparison optimization module; wherein the vascular risk analysis module, the tissue edema analysis module, the swine fever risk assessment module, and the comparison optimization module are connected in communication; The blood flow of limb blood vessels is analyzed through the vascular risk analysis module to assess the degree of vascular risk; the tissue edema analysis module is used to analyze the tissue edema of limb parts and assess the degree of tissue edema; the swine fever risk assessment module combines the vascular risk degree with the tissue edema degree to comprehensively assess the suspicion of swine fever and obtain the image detection results of swine fever virus, thereby providing auxiliary diagnosis for swine fever virus detection; the comparison and optimization module obtains the biological detection results of swine fever virus, and feeds them back to the image analysis model for comparison with the image detection results to determine the detection accuracy of the image analysis model, thereby continuously optimizing the image analysis model; Among them, the biological test results of the classical swine fever virus are diagnosed and obtained through existing pathological detection methods, and the biological test results are remotely transmitted to the data processing cloud. The pathological detection method is to obtain pathological samples of live pigs, such as blood samples, by regular sampling, and then conduct biological experiments on the pathological samples to determine whether the samples contain classical swine fever virus; for example, enzyme-linked immunosorbent assay, through antigen-antibody reaction to detect whether there are specific antibodies or antigens against classical swine fever virus in the serum; The specific construction process of the image analysis model is as follows: the image analysis model includes a vascular analysis sub-model, a tissue analysis sub-model and a comprehensive analysis sub-model; Sa, establish a vascular analysis sub-model through the vascular risk analysis module, analyze the blood flow of limb blood vessels, and evaluate the degree of vascular risk. The specific process of establishing the vascular analysis sub-model is as follows: Sa-1, input pig limb images to the vascular analysis sub-model; Extract N0 pixels of the pig limb image and collect grayscale, and mark the grayscale value of any pixel i as HDi; Set the grayscale interval Qh of the blood vessel, extract the pixel points whose grayscale values ​​HDi are in the grayscale interval Qh, and integrate them into a blood vessel feature point set, and mark the area where the blood vessel feature point set is located as a blood vessel area; The vascular area is divided into N1 characteristic regions and the regional state analysis is performed. Any characteristic region is marked as J, and the grayscale mean η1 of n0 pixels in characteristic region J is calculated: ; Set the grayscale interval Qη of regional infarction. When the grayscale mean η1 of the feature region J is within the grayscale interval Qη, the feature region J is judged to be vascular infarction and necrosis, and the feature region is marked as the infarction region; otherwise, the feature region J is judged to be vascular blood flow, and the feature region is marked as the blood flow region; Sa-2, conduct targeted analysis on the infarct area and blood flow area respectively. The specific process is as follows: Sa-201, analyze the infarct area: mark the number of infarct areas as n1, mark the area of ​​any infarct area as Sg, mark the area of ​​the vascular area as Sp, then cumulatively calculate the proportion of all infarct areas in the vascular area and mark it as the infarct proportion η2: ; Sa-202, blood flow area analysis: extract the pig limb images corresponding to N2 image acquisition cycles Tp, and construct a dynamic curve S0 between the grayscale mean η1 of the blood flow area and the image acquisition cycle Tp; The blood pumping cycle Tb and the blood pumping intensity Db are calculated by the dynamic curve S0. The specific process is as follows: the peak points and valley points of the dynamic curve S0 are obtained by calculating the slope values ​​of n2 points of the dynamic curve S0, and the number of peak points and valley points are marked as N3, and any group of peak points and valley points is marked as a peak-valley pair; By calculating the horizontal coordinate difference between two adjacent peak points, the peak point cycle length is obtained and marked as the blood pumping cycle Tb; and by calculating the vertical coordinate difference between the adjacent peak point and valley point, the peak-valley floating difference e1 is obtained, and then the vertical coordinate mean between the adjacent peak point and valley point is measured to obtain the peak-valley grayscale mean e2; Then, the blood pumping intensity Db is analyzed by combining the peak-to-valley floating difference e1 and the peak-to-valley grayscale mean e2 of the peak-to-valley pairs of group N3: ; Among them, α1 and α2 are weight factor coefficients of the peak-to-valley floating difference e1 and the peak-to-valley grayscale mean e2, respectively, and α1 and α2 are both greater than 0. The weight factor coefficients are preset and obtained after a large amount of experimental data is measured. When the peak-to-valley floating difference e1 and the peak-to-valley grayscale mean e2 are higher, the blood pumping intensity Db is higher, indicating that the blood volume difference between the contraction and extension states of the blood vessels in the blood pumping cycle Tb is large and the overall level is high, indicating that the blood state in the blood vessels is relatively active, and may be in a congested state, and the blood flow intensity is higher; Sa-3, by combining the infarction ratio η2, pumping cycle Tb and pumping intensity Db, the vascular risk assessment coefficient Xxg is obtained: ; Among them, the blood flow state coefficient is obtained by combining the blood pumping cycle Tb and the blood pumping intensity Db, and ω1 and ω2 are set as the weight factor coefficients of the infarction ratio η2 and the blood flow state coefficient respectively, and ω1 and ω2 are both greater than 0; when the blood pumping cycle Tb is lower and the blood pumping intensity Db is higher, the blood flow state coefficient is higher, indicating that the blood flow state is stronger. At the same time, when the infarction ratio η2 is higher, it means that the vascular infarction or necrosis is more serious, and the vascular risk assessment coefficient Xxg is higher, indicating that the comprehensive risk of vascular infarction state and blood flow state is higher, which will make the symptoms of swine fever more severe and the suspicion of swine fever virus infection higher; Sb, establish a tissue analysis sub-model through the tissue edema analysis module, analyze the tissue edema of the limbs, and evaluate the degree of tissue edema. The specific process of establishing the tissue analysis sub-model is as follows: Sb-1, extract the pixel points whose grayscale values ​​HDi of the pig limb image are not located in the grayscale interval Qh, that is, all the pixel points outside the vascular feature point set, and mark the area where they are located as the tissue area, that is, the non-vascular area; Sb-2, divide the tissue area into M1 characteristic regions and perform regional state analysis, mark any characteristic region as R, and calculate the grayscale mean φ1 of m0 pixels in the characteristic region R; Sb-201, set the grayscale interval Qs of regional edema, and if the grayscale mean φ1 of the extracted feature region R is within the grayscale interval Qs, it is determined that the feature region R is in a tissue edema state, and the feature region is marked as an edema region; otherwise, the feature region is determined to be normal; Sb-202, mark the number of edema areas in the pig limb image as m1, mark the area of ​​any edema area as Sz, then cumulatively calculate the proportion of all edema areas in the tissue area and mark it as edema proportion φ2; Extract the pig limb images corresponding to N2 image acquisition cycles Tp, and obtain the edema percentage of the pig limb images corresponding to N2 image acquisition cycles Tp; The two adjacent image acquisition periods are marked as Tp1 and Tp2 respectively, the edema proportion corresponding to the image acquisition period Tp1 is marked as φ2a, and the edema proportion corresponding to the image acquisition period Tp2 is marked as φ2b; Sb-203, calculate the growth rate ψ of the edema proportion φ2 corresponding to two adjacent image acquisition cycles: , and through the pig limb images corresponding to N2 image acquisition cycles Tp, the edema percentage growth rate corresponding to two image acquisition cycles is calculated in turn, so as to obtain the M2 edema percentage growth rate ψ, where, ; The average growth rate ψ of the M2 edema proportions is used to calculate the average growth rate Ψr of edema: , and then calculate the standard deviation to obtain the edema variation coefficient σr: ; Sb-3, by combining the average edema increase rate Ψr and the edema variation coefficient σr, the tissue risk assessment coefficient Xzh is obtained: ; Among them, μ1 and μ2 are the weight factor coefficients of the average increase rate of edema Ψr and the coefficient of edema variation σr, respectively, and both μ1 and μ2 are greater than 0; when the average increase rate of edema Ψr and the coefficient of edema variation σr are higher, the tissue risk assessment coefficient Xzh is higher, indicating that the edema growth trend is more obvious and the floating change degree is higher, the tissue edema risk degree is higher, which will make the symptoms of swine fever more severe and the suspicion of swine fever virus infection higher; Sc, a comprehensive analysis sub-model is established through the swine fever risk assessment module, the vascular risk level is combined with the tissue edema level, and the suspected degree of swine fever is comprehensively assessed to obtain the image detection results of swine fever virus, so as to assist in the diagnosis of swine fever virus detection. The specific process of establishing the comprehensive analysis sub-model is as follows: Sc-1, construct the change curve s1 between the vascular risk assessment coefficient Xxg and the image acquisition period Tp; Construct a variation curve s2 between the tissue risk assessment coefficient Xzh and the image acquisition period Tp; Synchronously extract M3 points of the change curve s1 and the change curve s2 to establish the risk parameter matrix V: ; Sc-2, by finding the mean of the column vector of the risk parameter matrix V, the overall risk index va is obtained, and then the variance of the column vector is measured to obtain the risk volatility index vb, and then the risk factor υ is obtained by combining the overall risk index va and the risk volatility index vb; Among them, the overall risk index va1 of the vascular risk assessment coefficient Xxg is: ; Risk volatility index vb1 of vascular risk assessment coefficient Xxg: ; Then the risk factor υ1 of the vascular risk assessment coefficient Xxg is: , set ρ １１ and ρ １２ are the weight factor coefficients of the overall risk index va1 and the risk volatility index vb1, and ρ １１ and ρ １２ All are greater than 0; The overall risk index va2 of the organizational risk assessment coefficient Xzh: ; Risk volatility index vb2 of organizational risk assessment coefficient Xzh: ; Then the risk factor υ2 of the organizational risk assessment coefficient Xzh is: , set ρ ２１ and ρ ２２ are the weight factor coefficients of the overall risk index va2 and the risk volatility index vb2, and ρ ２１ and ρ ２２ All are greater than 0; Sc-3, by combining the risk factor υ1 of the vascular risk assessment coefficient Xxg and the risk factor υ2 of the tissue risk assessment coefficient Xzh, the suspected reference index Zw of swine fever is obtained: ; Among them, ε1 and ε2 are set as the weight factor coefficients of risk factor υ1 and risk factor υ2 respectively, and ε1 and ε2 are both greater than 0; when the risk factor υ1 of the vascular risk assessment coefficient Xxg and the risk factor υ2 of the tissue risk assessment coefficient Xzh are higher, the swine fever suspected reference index Zw is higher, and thus the swine fever suspected degree is higher; Set the determination interval Qw of the suspected reference index Zw of classical swine fever. When the suspected reference index Zw of classical swine fever is within the determination interval Qw, the image detection result is determined to be infected with classical swine fever virus; otherwise, the image detection result is determined to be not infected with classical swine fever virus, and the image detection result and the suspected reference index Zw of classical swine fever are sent to the pig breeding management backend for professional management personnel to assist in the diagnosis of classical swine fever virus detection; Sd, obtains the biological detection results of classical swine fever virus through the comparison optimization module, and feeds it back to the image analysis model for comparison with the image detection results, determines the detection accuracy of the image analysis model, and continuously optimizes the image analysis model. The specific optimization process of the image analysis model is as follows: Sd-1, mark the total pig population as ZT, set the sampling ratio q%, and obtain the pig sample size YB: , select YB live pig samples by random sampling, and perform biological tests on the live pig samples to determine whether they are infected with swine fever virus, and set the biological test value Ja; If the biological test result is infection with swine fever virus, the biological test value Ja=1; If the biological test result is not infected with swine fever virus, the biological test value Ja=0; Sd-2, set the image detection value Pa; When the image detection result is infected with swine fever virus, the image detection value Pa is set to 1; When the image detection result is not infected with swine fever virus, the image detection value Pa is set to 0; Sd-3, compare the biological detection value Ja and the image detection value Pa to obtain the detection deviation coefficient JP: ; When the biological detection value Ja and the image detection value Pa are different, it means that there is a deviation in the image detection result; on the contrary, when the biological detection value Ja and the image detection value Pa are the same, it means that the image detection result is correct; by comparing the detection results of YB pig samples, the detection accuracy of the image analysis model is cumulatively measured and evaluated; Sd-4, set the evaluation interval of the detection deviation coefficient JP, and evaluate the detection accuracy of the image analysis model through interval comparison; There are u evaluation intervals of the preset detection deviation coefficient JP, and any evaluation interval is marked as Qu. When the detection deviation coefficient JP is in the evaluation interval Qu, the detection accuracy of the evaluation image analysis model is u level. When the detection deviation coefficient JP is higher, the detection accuracy of the evaluation image analysis model is lower. When the detection accuracy level of the image analysis model is u, the correction amplitude Fu% of the image analysis model is set, and the preset parameters of the image analysis model are adjusted by the correction amplitude Fu%; Among them, the prediction parameters of the image analysis model include weight factor coefficients ε1 and ε2. When the image detection value Pa=1 and the biological detection value Ja=0, it means that the image detection value Pa is too large, which means that the suspected reference index Zw of swine fever is too large. In this case, the weight factor coefficients ε1 and ε2 need to be reduced. The adjusted weight factor coefficients are marked as ε1t and ε2t respectively. , , by correcting the preset parameters with corresponding amplitudes, the model fitting accuracy is continuously improved, thereby improving the detection accuracy of the image analysis model; S3, the disease monitoring unit conducts long-term monitoring of the classical swine fever virus: by regularly monitoring the image detection results of the classical swine fever virus, analyzing the long-term development status of the classical swine fever virus, assessing the health risk level of sick pigs, generating a health management plan for sick pigs and providing corresponding risk warnings. The specific process is as follows: When both the image detection result and the biological detection result of classical swine fever virus are infected with classical swine fever virus, the live pig sample is marked as a sick pig, the number of sick pigs is marked as BZ, the classical swine fever suspected reference index Zw corresponding to the image detection result of the sick pig is extracted, and the image detection result of classical swine fever virus is monitored regularly to analyze the long-term development status of classical swine fever virus in the sick pig; Set the virus monitoring period Td, take the swine fever suspected reference index Zw as the ordinate and the virus monitoring period Td as the abscissa, construct the change curve sd between the swine fever suspected reference index Zw and the virus monitoring period Td, extract n3 points of the change curve sd and their coordinates, calculate and obtain the slope value corresponding to the n3 points, and mark the slope value of any point as Kpo; The slope Kpo is calculated by referring to the growth rate ψ calculation formula, and is calculated by the coordinates of adjacent points of the change curve sd; The average value is calculated by the slope values ​​Kpo of n3 points to obtain the curve increase rate Ψs of the change curve sd, and then the standard deviation is calculated by the slope values ​​Kpo of n3 points to obtain the curvature fluctuation coefficient σs of the change curve sd; When the slope value Kpo of any point is 0, the point is determined to be a marked point, and the slope value of the point corresponding to the next period after the marked point is obtained and marked as Kpj. If the slope value Kpj is greater than 0, it means that there is an upward trend after the marked point, and the marked point is determined to be a valley point; if the slope value Kpj is less than 0, it means that there is a downward trend after the marked point, and the marked point is determined to be a peak point; Mark the nearest annotation point to the end of the change curve sd as Yz, extract the curve segment between the annotation point Yz and the end point of the change curve sd and mark it as sz, mark the coordinates of the annotation point Yz as (Xz, Yz), mark the coordinates of the end point of the change curve sd as (Xm, Ym), and analyze the drop value Lz and change rate Kz of the curve segment sz; Among them, the drop value Lz is: ; Change rate Kz: ; If the marked point is a peak point, the peak point is marked as POf, the ordinate of the peak point POf is marked as Yf, the curve segment between the peak point POf and the end point of the change curve sd is extracted and marked as sf, then the change rate Kz of the curve segment sf is a negative value, indicating that the curve segment sf is in a downward trend, indicating that the swine fever suspected reference index Zw is in a downward state, which indicates that the swine fever virus of the sick pig is decreasing and the health level is recovering; when the drop value Lz of the curve segment sf is higher, it means that the degree of recovery of the health status of the sick pig is higher; If the marked point is a valley point, the valley point is marked as POg, the ordinate of the valley point POg is marked as Yg, the curve segment between the valley point POg and the end point of the change curve sd is extracted and marked as sg, then the change rate Kz of the curve segment sg is a positive value, indicating that the curve segment sg is on an upward trend, indicating that the swine fever suspected reference index Zw is on the rise, which indicates that the swine fever virus of the sick pig is on the rise and the physical condition is deteriorating; when the drop value Lz of the curve segment sg is higher, it means that the degree of deterioration of the physical condition of the sick pig is higher; The health risk index Zjk of sick pigs is obtained by combining the curve increase rate Ψs and curvature fluctuation coefficient σs of the change curve sd, as well as the drop value Lz and change rate Kz of the curve segment sz: ; Among them, the overall level of health risk is analyzed by combining the curve increase rate Ψs and the curvature fluctuation coefficient σs of the change curve sd. When the curve increase rate Ψs and the curvature fluctuation coefficient σs of the change curve sd are higher, it means that the overall level of the risk of classical swine fever virus is high and the fluctuation amplitude is large, indicating that the virus state is active and unstable; then the current change of health risk is analyzed by the drop value Lz and the change rate Kz of the curve segment sz; when the drop value Lz and the change rate Kz of the curve segment sz are positive and higher, it means that the current risk change of classical swine fever virus is gradually deteriorating and the degree of deterioration is high; when the drop value Lz and the change rate Kz of the curve segment sz are negative and the absolute value is higher, it means that the current risk change of classical swine fever virus is gradually recovering and the degree of recovery is high; set δ1 and δ2 as the weight factor coefficients of the overall risk level and the current risk change, and δ1 and δ2 are both greater than 0; Set the evaluation interval of the health risk index Zjk of sick pigs, and evaluate the health risk level of sick pigs by comparing the intervals. The higher the health risk index Zjk of sick pigs, the higher the health risk level of sick pigs; The evaluation interval of the health risk index Zjk of the sick pig is preset to be [Qj1, Qj2]. When the health risk index Zjk of the sick pig is lower than Qj1, the health risk level of the sick pig is judged to be mild; when the health risk index Zjk of the sick pig is in the evaluation interval [Qj1, Qj2], the health risk level of the sick pig is judged to be moderate; when the health risk index Zjk of the sick pig is higher than Qj2, the health risk level of the sick pig is judged to be severe; Generate corresponding sick pig health management plans and provide risk warnings according to the health risk level of sick pigs; When the health risk level of sick pigs is mild, a Level I sick pig health management plan and a No. 1 risk warning signal are generated; when the health risk level of sick pigs is moderate, a Level II sick pig health management plan and a No. 2 risk warning signal are generated; when the health risk level of sick pigs is severe, a Level III sick pig health management plan and a No. 3 risk warning signal are generated; Among them, the health management plan for sick pigs is preset by breeding managers according to the health risk level of sick pigs. For example, the health management plan for sick pigs at level I uses enhanced immunization measures, conducts secondary or booster vaccinations, enhances immune response, provides easily digestible and nutritious feed, and improves the body resistance of sick pigs; the health management plan for sick pigs at level II uses appropriate drugs for symptomatic treatment based on the doctor's diagnosis, and uses drugs to relieve the discomfort symptoms of sick pigs; the health management plan for sick pigs at level III is a comprehensive implementation of the health management plan for sick pigs at level I and level II; In this way, different degrees of treatment and control can be carried out on sick pigs that may be infected with the swine fever virus during pig breeding, and the efficiency of sick pig health management can be improved, thereby improving the targeted treatment effect of sick pigs, inhibiting the further deterioration and spread of the swine fever virus, and reducing pig breeding losses and management costs; S4, the early warning management unit conducts a global analysis of the virus risk: by conducting an overall analysis of the virus risk level, assessing the infection status of the swine fever virus in the pig breeding base, generating a pig breeding management plan and conducting corresponding early warning management. The specific process is as follows: By conducting an overall analysis of the virus risk level of all pigs in the pig breeding base, the classical swine fever virus infection index Zgr is obtained: ; The higher the proportion of the number of sick pigs BZ detected in the pig sample YB, the higher the estimated number of sick pigs in the total pig sample ZT. The higher the health risk index Zjk of sick pigs, the higher the deterioration of the classical swine fever virus. As a result, the higher the classical swine fever virus infection index Zgr, the more serious the classical swine fever virus infection status of the pig breeding base is. Set the evaluation interval of the classical swine fever virus infection index Zgr, and evaluate the classical swine fever virus infection degree of the pig breeding base by comparing the intervals. The higher the classical swine fever virus infection index Zgr, the higher the classical swine fever virus infection degree of the pig breeding base; The evaluation interval of the classical swine fever virus infection index Zgr is preset to [Qr1, Qr2]. When the classical swine fever virus infection index Zgr is lower than Qr1, the classical swine fever virus infection degree is judged to be mild; when the classical swine fever virus infection index Zgr is in the evaluation interval [Qr1, Qr2], the classical swine fever virus infection degree is judged to be moderate; when the classical swine fever virus infection index Zgr is higher than Qr2, the classical swine fever virus infection degree is judged to be severe; According to the infection status of swine fever virus in the swine breeding base, generate corresponding swine breeding management plan and carry out early warning management; When the infection level of the swine fever virus is mild, a Level I pig breeding management plan and a No. 1 early warning management signal are generated; when the infection level of the swine fever virus is moderate, a Level II pig breeding management plan and a No. 2 early warning management signal are generated; when the infection level of the swine fever virus is severe, a Level III pig breeding management plan and a No. 3 early warning management signal are generated; Among them, the pig breeding management plan is preset by breeding management personnel according to the infection status of the swine fever virus in the pig breeding base. For example, the Level I pig breeding management plan takes preventive measures, including regular swine fever vaccination of the pig herd to enhance immunity, regular inspection of the health of the pig herd, early detection of suspected cases, maintenance of breeding environment hygiene, and regular disinfection of feed, utensils and vehicles; the Level II pig breeding management plan is for emergency response. For sudden outbreaks, a clear emergency handling process is formulated, suspected infected pigs are isolated and observed to prevent transmission, and the breeding area is thoroughly cleaned and disinfected to eliminate potential viruses; the Level III pig breeding management plan is for disease treatment, through drug treatment to detect its health status, control the degree of deterioration of the disease, and harmlessly handle dead pigs, and strictly handle dead pigs in accordance with regulations, such as incineration and burial, to avoid the spread of the virus; This will enable unified control of the overall classical swine fever virus infection status of the pig breeding base to varying degrees, improve the early warning management efficiency of the pig breeding base, thereby inhibiting the further deterioration and spread of the classical swine fever virus, and reducing pig breeding losses and management costs.

[0016] In summary, the present invention remotely collects images of pig limbs and sends them to the cloud for data processing, thereby reducing the demand for manual sampling, and constructs an image analysis model to evaluate the suspected degree of swine fever virus infection, thereby improving the efficiency of virus detection. The accuracy of the image analysis model is compared and tested by sampling biological detection, thereby continuously optimizing the image analysis model, improving the accuracy of virus detection, significantly reducing the trend of virus infection spread, and effectively curbing the continued deterioration of the swine fever epidemic. The present invention collects images of pig limbs through an image acquisition unit, and constructs an image analysis model on a data processing cloud to analyze the images of pig limbs, comprehensively assesses the suspected risk of classical swine fever from the two perspectives of vascular risk and tissue edema, thereby performing auxiliary diagnosis for classical swine fever virus detection; and then performs long-term monitoring of classical swine fever virus through a disease monitoring unit, analyzes the deterioration or recovery of the condition of sick pigs, generates a health management plan for sick pigs, thereby performing corresponding health management according to the health risk degree of sick pigs, and improving the targeted treatment effect of sick pigs; The present invention conducts a global analysis of virus risks through an early warning management unit and generates a pig breeding management plan, thereby performing corresponding overall control on the infection level of classical swine fever virus in the pig breeding base, improving the early warning management efficiency of the pig breeding base, and reducing pig breeding losses and management costs.

[0017] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technical personnel in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0018] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. The swine fever virus detection system based on cloud analysis is characterized by: It includes an image acquisition unit, a data processing cloud, a disease monitoring unit and an early warning management unit, wherein the image acquisition unit, the data processing cloud, the disease monitoring unit and the early warning management unit are communicatively connected; The image acquisition unit is used to collect images of pig limbs: the image acquisition period Tp is set, and the wearable device is bound to the pig limbs, and an infrared image acquisition device is set in the wearable device. The infrared image acquisition device is used to regularly acquire images of pig limbs, and then the images of pig limbs are transmitted to the data processing cloud through remote signals; The data processing cloud is used to build image analysis models to analyze pig limb images: the data processing cloud includes a vascular risk analysis module, a tissue edema analysis module, a swine fever risk assessment module, and a comparison optimization module; The blood flow of limb blood vessels is analyzed through the vascular risk analysis module to assess the degree of vascular risk; the tissue edema analysis module is used to analyze the tissue edema of limb parts and assess the degree of tissue edema; the swine fever risk assessment module combines the vascular risk degree with the tissue edema degree to comprehensively assess the suspicion of swine fever and obtain the image detection results of swine fever virus, thereby providing auxiliary diagnosis for swine fever virus detection; the comparison and optimization module obtains the biological detection results of swine fever virus, and feeds them back to the image analysis model for comparison with the image detection results to determine the detection accuracy of the image analysis model, thereby continuously optimizing the image analysis model; The disease monitoring unit is used for long-term monitoring of the classical swine fever virus: by regularly monitoring the image detection results of the classical swine fever virus, analyzing the long-term development status of the classical swine fever virus, assessing the health risk level of sick pigs, generating a health management plan for sick pigs and providing corresponding risk warnings; The early warning management unit is used to conduct a global analysis of virus risks: by conducting an overall analysis of the virus risk level, evaluating the infection status of the swine fever virus in the pig breeding base, generating a pig breeding management plan and conducting corresponding early warning management.

2. The classical swine fever virus detection system based on cloud analysis according to claim 1, characterized in that: The specific construction process of the image analysis model is as follows: The image analysis model includes a vascular analysis sub-model, a tissue analysis sub-model and a comprehensive analysis sub-model; Sa, establishes a vascular analysis sub-model through the vascular risk analysis module to analyze the blood flow of limb blood vessels and assess the degree of vascular risk; Sb, establishes a tissue analysis sub-model through the tissue edema analysis module to analyze the tissue edema of the limbs and evaluate the degree of tissue edema; Sc, a comprehensive analysis sub-model is established through the swine fever risk assessment module, which combines the vascular risk level with the tissue edema level, comprehensively assesses the suspicion of swine fever, and obtains the image detection results of swine fever virus, thereby assisting the diagnosis of swine fever virus detection; Sd, obtains the biological detection results of classical swine fever virus through the comparison optimization module, and feeds it back to the image analysis model for comparison with the image detection results to determine the detection accuracy of the image analysis model, thereby continuously optimizing the image analysis model.

3. The classical swine fever virus detection system based on cloud analysis according to claim 2, characterized in that: The specific process of establishing the vascular analysis sub-model is as follows: Sa-1, input pig limb images to the vascular analysis sub-model; Extract N0 pixels of the pig limb image and collect grayscale, and mark the grayscale value of any pixel i as HDi; Set the grayscale interval Qh of the blood vessel, extract the pixel points whose grayscale values ​​HDi are in the grayscale interval Qh, and integrate them into a blood vessel feature point set, and mark the area where the blood vessel feature point set is located as a blood vessel area; The blood vessel area is divided into N1 characteristic regions and the regional state analysis is performed. Any characteristic region is marked as J, and the grayscale mean η1 of n0 pixels in the characteristic region J is calculated; Set the grayscale interval Qη of regional infarction. When the grayscale mean η1 of the feature region J is within the grayscale interval Qη, the feature region is marked as the infarction region; otherwise, the feature region is marked as the blood flow region. Sa-2, conduct targeted analysis on the infarct area and blood flow area respectively. The specific process is as follows: Sa-201, analysis of infarct area: by calculating the proportion of the total infarct area in the vascular area, the infarct proportion η2 was obtained; Sa-202, blood flow area analysis: extract the pig limb images corresponding to N2 image acquisition cycles Tp, and construct a dynamic curve S0 between the grayscale mean η1 of the blood flow area and the image acquisition cycle Tp, and obtain the blood pumping cycle Tb and blood pumping intensity Db through the dynamic curve S0; Sa-3, by combining the infarction proportion η2, pumping cycle Tb and pumping intensity Db, the vascular risk assessment coefficient Xxg is obtained.

4. The classical swine fever virus detection system based on cloud analysis according to claim 3, characterized in that: The specific process of establishing the organizational analysis sub-model is as follows: Sb-1, extracting the pixel points whose grayscale values ​​HDi of the pig limb image are not located in the grayscale interval Qh, and marking the area where the pixel points are located as the tissue area; Sb-2, divide the tissue area into M1 characteristic regions and perform regional state analysis, mark any characteristic region as R, and calculate the grayscale mean φ1 of m0 pixels in the characteristic region R; Sb-201, set the grayscale interval Qs of regional edema, and if the grayscale mean φ1 of the extracted feature region R is in the grayscale interval Qs, then mark the feature region as an edema region; Sb-202, by cumulatively measuring the proportion of all edema areas in the tissue area, obtaining the edema proportion φ2, extracting the pig limb images corresponding to N2 image acquisition cycles Tp, and obtaining the edema proportion of the pig limb images corresponding to N2 image acquisition cycles Tp; Sb-203, obtain the edema proportion growth rate ψ through the edema proportion φ2 corresponding to two adjacent image acquisition cycles, and then obtain the average edema growth rate Ψr and the edema variation coefficient σr; Sb-3, by combining the average edema increase rate Ψr and the edema variation coefficient σr, the tissue risk assessment coefficient Xzh is obtained.

5. The classical swine fever virus detection system based on cloud analysis according to claim 4, characterized in that: The specific process of establishing the comprehensive analysis sub-model is as follows: Sc-1, constructing a change curve s1 between the vascular risk assessment coefficient Xxg and the image acquisition period Tp, and a change curve s2 between the tissue risk assessment coefficient Xzh and the image acquisition period Tp; Synchronously extract M3 points of the change curve s1 and the change curve s2 to establish the risk parameter matrix V; Sc-2, by obtaining the overall risk index va and the risk fluctuation index vb from the risk parameter matrix V, and then obtaining the risk factor υ, the risk factors of the vascular risk assessment coefficient Xxg and the tissue risk assessment coefficient Xzh are marked as υ1 and υ2 respectively; Sc-3, by combining the risk factor υ1 of the vascular risk assessment coefficient Xxg and the risk factor υ2 of the tissue risk assessment coefficient Xzh, the suspected reference index Zw of swine fever is obtained; A determination interval Qw of the swine fever suspected reference index Zw is set. When the swine fever suspected reference index Zw is within the determination interval Qw, the image detection result is determined to be infected with swine fever virus; otherwise, the image detection result is determined to be not infected with swine fever virus.

6. The classical swine fever virus detection system based on cloud analysis according to claim 5, characterized in that: The specific optimization process of the image analysis model is: Sd-1, mark the total pig population as ZT, set the sampling ratio q%, obtain the pig sample volume YB, select YB pig samples by random sampling, perform biological tests on the pig samples, obtain the biological test results, set the biological test value Ja, and assign the biological test value Ja according to the biological test results; Sd-2, set the image detection value Pa, and assign the image detection value Pa according to the image detection result; Sd-3, compare the biological detection value Ja and the image detection value Pa to obtain the detection deviation coefficient JP; Sd-4, set the evaluation interval of the detection deviation coefficient JP, evaluate the detection accuracy of the image analysis model through interval comparison, and then set the correction amplitude of the image analysis model to optimize and adjust the image analysis model.

7. The classical swine fever virus detection system based on cloud analysis according to claim 6, characterized in that: The specific process of long-term monitoring of classical swine fever virus is as follows: When both the image detection result and the biological detection result of the classical swine fever virus are infected with the classical swine fever virus, the live pig sample is marked as a sick pig, the number of sick pigs is marked as BZ, and the classical swine fever suspected reference index Zw corresponding to the image detection result of the sick pig is extracted; Set the virus monitoring period Td, construct the change curve sd between the suspected reference index Zw of swine fever and the virus monitoring period Td, extract n3 points of the change curve sd and their coordinates, calculate and obtain the slope value corresponding to the n3 points, and obtain the curve increase rate Ψs and curvature fluctuation coefficient σs of the change curve sd; When the slope value Kpo of any point is 0, the point is determined to be a marked point, and the curve segment sz between the marked point Yz and the end point of the change curve sd is extracted to obtain the drop value Lz and the change rate Kz of the curve segment sz; Then, the health risk index Zjk of sick pigs is comprehensively obtained, and the evaluation interval of the health risk index Zjk of sick pigs is set. The health risk degree of sick pigs is evaluated through interval comparison, and the corresponding sick pig health management plan is generated and risk warnings are issued.

8. The classical swine fever virus detection system based on cloud analysis according to claim 7, characterized in that: The specific process of conducting a global analysis of virus risks is as follows: The estimated number of sick pigs in the total pig volume ZT is obtained by measuring the proportion of the number of sick pigs BZ detected in the pig sample volume YB. Combined with the health risk index Zjk of sick pigs, the virus risk level of all pigs in the pig breeding base is analyzed as a whole to obtain the swine fever virus infection index Zgr; Set the evaluation range of the classical swine fever virus infection index Zgr, evaluate the degree of classical swine fever virus infection in the pig breeding base through interval comparison, generate the corresponding pig breeding management plan and conduct early warning management.

Images