A method and system for silkworm breeding and management based on multi-source data

Abstract

This invention discloses a silkworm breeding and management method and system based on multi-source data, belonging to the field of aquaculture management technology. This invention assesses the abnormal state of mulberry leaf diseases through mulberry leaf images, and adjusts environmental parameters in the larval stage based on the overall age of the silkworm and the abnormal state of mulberry leaf diseases. In the cocooning stage, environmental parameters are adjusted according to the abnormal state of the silkworm. The adjustment of environmental parameters is carried out according to different situations. The different growth cycles of the silkworm are used as the horizontal dimension and the abnormal state of mulberry leaf diseases is introduced as the vertical dimension. The adjustment of environmental parameters based on multiple dimensions improves the refinement and accuracy of silkworm breeding and management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of breeding management, and particularly relates to a silkworm cultivation management method and system based on multi-source data. BACKGROUND

[0002] Silkworm cultivation has multiple effects in economy, agriculture, technology and culture, such as that cocoon silk of silkworm is the main source of natural silk, which is used to produce silk fabrics, and silk protein (silk fibroin) is used to manufacture surgical sutures, artificial skin, drug release carriers and other biomedical materials, so that high-quality silkworm cultivation management is very important.

[0003] However, the existing silkworm cultivation management is basically managed by relying on experience through artificial observation, which is not only low in efficiency, but also lacks precision and refinement. SUMMARY

[0004] To solve the technical problems in the prior art, the present application provides a silkworm cultivation management method based on multi-source data, comprising the following steps:

[0005] various environment parameters in the silkworm room are acquired in real time through a sensor module, the overall period of the silkworm is evaluated through multiple first images, the overall period of the silkworm comprises a larva period and a cocooning period, multiple mulberry leaf images are randomly acquired, and the abnormal state of the mulberry leaf disease is evaluated through the mulberry leaf images.

[0006] If the overall period of the silkworm is the larva period, the following steps are performed:

[0007] multiple second silkworm images of silkworm individuals in the larva period are randomly acquired;

[0008] the overall age period of the silkworm is evaluated according to the second silkworm images;

[0009] the environment parameters are adjusted in combination with the overall age period of the silkworm and the abnormal state of the mulberry leaf disease;

[0010] If the overall period of the silkworm is the cocooning period, the following steps are performed:

[0011] multiple third silkworm images of silkworm individuals in the cocooning period are randomly acquired;

[0012] the overall abnormal state of the silkworm is evaluated according to the third silkworm images;

[0013] the environment parameters are adjusted according to the abnormal state of the silkworm.

[0014] Further, when the first image is a silkworm room image, the overall period of the silkworm is evaluated through the multiple first images, and specifically:

[0015] The first deep learning model is pre-trained, and whether a cluster device exists in the silkworm room image is identified. If the silkworm room image exists the cluster device, the silkworm room image is regarded as a first silkworm room image. If the ratio of the number of the first silkworm room image to the number of the silkworm room image is greater than a first preset threshold, the evaluation result of the period in which the silkworm as a whole is located is the cocooning period, otherwise, the evaluation result of the period in which the silkworm as a whole is located is the larva period.

[0016] Further, when the first image is a silkworm image, the period in which the silkworm as a whole is located is evaluated by using the plurality of first images, and specifically:

[0017] The silkworm features of each silkworm in the silkworm image are extracted, including morphological features, texture features and color features. The period in which each silkworm is located is identified according to the silkworm features by using a second deep learning model which is pre-trained. If the number of silkworms in the larva period is greater than the number of silkworms in the cocooning period in the silkworm image, the silkworm image is a larva period image, otherwise, the silkworm image is a cocooning period image. If the number of the larva period images is greater than the number of the cocooning period images, the evaluation result of the period in which the silkworm as a whole is located is the larva period, otherwise, the evaluation result of the period in which the silkworm as a whole is located is the cocooning period.

[0018] Further, the disease abnormality of the mulberry leaf is evaluated by using the mulberry leaf image, and specifically:

[0019] The mulberry leaf region is segmented in the mulberry leaf image, and the edge of each mulberry leaf is identified in the mulberry leaf region to obtain a plurality of mulberry sub-regions representing each mulberry leaf;

[0020] The pixel values of each pixel in each mulberry sub-region in three color channels of RGB are obtained, and the channel average of each color channel corresponding to each mulberry sub-region is calculated. For each color channel average, it is determined whether it is outside the corresponding preset normal value range. If yes, the channel average is an abnormal channel average.

[0021] The mulberry sub-region with the abnormal channel average is regarded as a first abnormal region. If the ratio of the number of the first abnormal region to the number of the mulberry sub-region is less than a preset ratio, the mulberry leaf state is evaluated as normal, otherwise, the following steps are continued to be executed.

[0022] Abnormal pixel points are obtained in each first abnormal region. The abnormal pixel point is a pixel point with an absolute value of the difference between the pixel value of one or more color channels and the abnormal channel average of the corresponding color channel in the first abnormal region being less than or equal to a preset difference value.

[0023] In each first abnormal region, consecutive adjacent abnormal pixel points are regarded as a second abnormal region, and the shape feature of each second abnormal region is obtained. The disease abnormality of the mulberry leaf is estimated according to the shape feature of each second abnormal region in each mulberry sub-region and the abnormal channel average of the existing color channel by using a third deep learning model which is pre-trained.

[0024] Further, the process of acquiring multiple environment parameters of the silkworm room by the sensor module further comprises temperature compensation correction of each sensor by a preset neural network model, and the neural network model is optimized by a white whale optimization algorithm on model hyperparameters in a training process.

[0025] Further, the evaluation of the overall age of the domestic silkworm is specifically:

[0026] Color features, texture features and body length features of the corresponding larva stage domestic silkworm individuals are extracted according to each second domestic silkworm image, and the age of the corresponding larva stage domestic silkworm individuals is analyzed, and finally the age with the largest number of larva stage domestic silkworm individuals is taken as the overall age of the domestic silkworm.

[0027] Further, the adjustment of the environment parameters in combination with the overall age of the domestic silkworm and the abnormal state of mulberry leaf disease is specifically:

[0028] A first adaptive range of each environment parameter is acquired according to the overall age of the domestic silkworm.

[0029] A second adaptive range of each environment parameter is acquired according to the abnormal state of mulberry leaf disease.

[0030] If the first adaptive range and the second adaptive range of the corresponding environment parameter have an intersection, the intersection is taken as a third adaptive range of the corresponding environment parameter, and the corresponding environment parameter is adjusted to be within the corresponding third adaptive range.

[0031] If the first adaptive range and the second adaptive range of the corresponding environment parameter have no intersection, a preset weight of the overall age of the domestic silkworm and a preset weight of the abnormal state of mulberry leaf disease are acquired, if the preset weight of the overall age of the domestic silkworm is greater than the preset weight of the abnormal state of mulberry leaf disease, the corresponding environment parameter is adjusted to be within the corresponding first adaptive range, otherwise it is adjusted to be within the corresponding second adaptive range.

[0032] Further, the evaluation of the overall abnormal state of the domestic silkworm according to each third domestic silkworm image is:

[0033] Color features, morphological features and cocoon silk color features of the domestic silkworm are extracted in the third domestic silkworm image, and a first abnormal type existing in the corresponding cocooning period domestic silkworm individual is analyzed.

[0034] In all third domestic silkworm images, the number of occurrences of each first abnormal type is counted, if the ratio of the number of occurrences of the first abnormal type to the number of third domestic silkworm images is greater than or equal to a preset number ratio, the corresponding first abnormal type is a second abnormal type, and a set of each second abnormal type is taken as the overall abnormal state of the domestic silkworm.

[0035] The application also provides a domestic silkworm cultivation management system based on multi-source data, comprising:

[0036] The sensor module acquires various environmental parameters in the silkworm rearing room in real time.

[0037] The first evaluation module assesses the overall life cycle of the silkworm using multiple first images, including the larval stage and the cocooning stage.

[0038] The second evaluation module randomly acquires multiple mulberry leaf images and evaluates the abnormal status of mulberry leaf diseases through these images.

[0039] The first adjustment module randomly acquires second images of multiple larval silkworm individuals when the overall silkworm cycle is in the larval stage, evaluates the overall age of the silkworm based on the second silkworm images, and adjusts environmental parameters in combination with the overall age of the silkworm and the abnormal state of mulberry leaf diseases.

[0040] The second adjustment module randomly acquires third images of multiple silkworm individuals in the cocooning stage when the overall silkworm cycle is in the cocooning stage. It then assesses the overall abnormal state of the silkworm based on each third silkworm image and adjusts the environmental parameters accordingly.

[0041] Furthermore, the assessment of abnormal mulberry leaf disease conditions using the mulberry leaf images specifically involves:

[0042] The mulberry leaf region is segmented in the mulberry leaf image, and the edge of each mulberry leaf is identified in the mulberry leaf region to obtain multiple mulberry leaf regions representing each mulberry leaf;

[0043] The pixel values ​​of the RGB three color channels of each pixel in each mulberry leaf area are obtained respectively. The channel mean of each color channel corresponding to each mulberry leaf area is calculated. For each color channel mean, it is determined whether it is outside the corresponding preset normal value range. If so, the channel mean is an abnormal channel mean.

[0044] The mulberry leaf area with abnormal channel mean is regarded as the first abnormal area. If the ratio of the number of first abnormal areas to the number of mulberry leaf areas is less than the preset ratio, the mulberry leaf condition is assessed as normal; otherwise, the following steps are continued.

[0045] Abnormal pixels are obtained in each first abnormal region. The abnormal pixel is a pixel whose absolute value of the difference between the pixel value of one or more color channels and the average abnormal channel value of the corresponding color channel in the first abnormal region is less than or equal to a preset difference.

[0046] In each of the first abnormal regions, consecutive adjacent abnormal pixels are taken as second abnormal regions to obtain the shape features of each second abnormal region. Then, a pre-trained third deep learning model is used to estimate the disease abnormalities in the mulberry leaves based on the shape features of each second abnormal region in each mulberry leaf region and the mean value of the abnormal channels in the existing color channels.

[0047] Compared with the prior art, the present application has the beneficial effects that:

[0048] The present application evaluates the abnormal state of mulberry leaf diseases through mulberry leaf images, and adjusts environmental parameters in the larva stage in combination with the overall age of the silkworm and the abnormal state of mulberry leaf diseases, adjusts environmental parameters in the cocooning stage according to the abnormal state of the silkworm, adjusts environmental parameters in different cases, takes different growth cycles of the silkworm as the horizontal dimension and introduces the abnormal state of mulberry leaf diseases as the vertical dimension, adjusts environmental parameters in combination with multiple dimensions, and improves the refinement and accuracy of silkworm cultivation management.

[0049] By segmenting the mulberry leaf area and recognizing the single leaf edge, the interference of complex background on disease identification is effectively excluded; by using the RGB three-channel mean, the multi-dimensional information of the color space is comprehensively utilized, the local continuous area analysis of abnormal pixel points is combined, single pixel noise interference is avoided, the robustness is enhanced, and the accuracy of disease abnormal analysis is improved. BRIEF DESCRIPTION OF DRAWINGS

[0050] The accompanying drawings, which are incorporated into and form part of the specification, illustrate an embodiment consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0052] Figure 1 is a flowchart of a silkworm cultivation management method based on multi-source data of the present application. DETAILED DESCRIPTION

[0053] The technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0054] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications also change accordingly.

[0055] In addition, the description related to "first", "second" and the like in the present application is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the same, or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions of various embodiments can be combined with each other, but it must be based on the realization of ordinary skilled in the art, when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, nor within the protection scope required by the present application.

[0056] Embodiment one

[0057] Reference Figure 1 As shown in the drawings, the present application provides a mulberry breeding management method based on multi-source data, which specifically includes the following steps:

[0058] S1, real-time acquisition of multiple environmental parameters in the silkworm room through the sensor module, evaluation of the whole period of the silkworm through multiple first images, the whole period of the silkworm including the larva period and the cocoon period; randomly acquire multiple mulberry leaf images, and evaluate the abnormal state of mulberry leaf disease through the mulberry leaf images;

[0059] If the whole period of the silkworm is the larva period, the following steps are executed:

[0060] S21, randomly acquiring multiple second silkworm images of silkworm individuals in the larva period;

[0061] S22, evaluating the whole age of the silkworm according to the second silkworm image;

[0062] S23, adjusting the environmental parameters in combination with the whole age of the silkworm and the abnormal state of mulberry leaf disease;

[0063] If the whole period of the silkworm is the cocoon period, the following steps are executed:

[0064] S31, randomly acquiring multiple third silkworm images of silkworm individuals in the cocoon period;

[0065] S32, evaluating the whole abnormal state of the silkworm according to each third silkworm image;

[0066] S33, adjusting the environmental parameters according to the abnormal state of the silkworm.

[0067] The life cycle of the silkworm is roughly divided into egg stage, larva stage, cocooning stage and adult stage; the larva stage is the core stage of rapid growth and nutrient accumulation of the silkworm, the silk gland of the silkworm is fully developed in the larva stage, and the synthesis amount and quality of the silk protein determine the silk length, silk amount and strength of the cocoon, and therefore the nutrient intake and growth environment of the larva stage are directly related to the development of the silk gland, and if the amount and quality of the mulberry leaves and the environmental conditions are insufficient in this stage, the healthy development of the larva will be affected, and then the subsequent silk spinning ability will be affected; the cocooning stage is the period of silk spinning and cocooning of the silkworm, and if the environmental conditions are improper, the silk spinning of the silkworm may be interrupted, the cocoon shape may be irregular or the silk quality may be fragile; therefore, the cultivation and management of the larva stage and the cocooning stage are particularly important in the cultivation of the silkworm.

[0068] Different periods of the silkworm have different cultivation conditions, and the larva stage and the cocooning stage are the same, each having different suitable cultivation and management strategies, and therefore it is necessary to analyze the period in which the silkworm is located, so as to better analyze the corresponding cultivation and management strategy, and the present application can efficiently and accurately evaluate the overall period in which the silkworm is located through image recognition.

[0069] It should be noted that the present application is only for the cultivation and management of the silkworm in the larva stage and the cocooning stage, that is, the evaluation of the overall period in which the silkworm is located is only in the larva stage and the cocooning stage.

[0070] In some embodiments, the first image in step S1 can be a silkworm chamber image, and when the first image is a silkworm chamber image, the evaluation of the overall period in which the silkworm is located through a plurality of first images is specifically:

[0071] Through the pre-trained first deep learning model, it is identified whether there is a cocooning tool in the silkworm chamber image, the silkworm chamber image with the cocooning tool is taken as a first silkworm chamber image, if the ratio of the number of the first silkworm chamber images to the number of the silkworm chamber images is greater than a first preset threshold, the evaluation result of the overall period in which the silkworm is located is the cocooning stage, otherwise it is the larva stage.

[0072] The first deep learning model is trained through a large number of various cocooning tool images as training samples, and can identify whether there is a cocooning tool in the image.

[0073] Upper cocooning is a key stage in the silkworm breeding process, in which mature silkworms are transferred to cocooning tools for cocooning, and the cocooning tool refers to a tool used by the silkworm for cocooning, such as a cocooning tool, that is, a support for silk spinning and cocooning, in the cocooning stage, the silkworm will find a suitable place to cocoon, so the cocooning tool is placed in the breeding process to help them, that is, the silkworm spins silk and cocoons on the cocooning tool, and the silkworm in the larva stage is still growing by eating mulberry leaves and does not need to cocoon, so it does not need a cocooning tool, therefore, in this embodiment, the present application analyzes the cocooning tool in the silkworm chamber, and then evaluates whether the overall period in which the silkworm in the silkworm chamber is located is in the cocooning stage.

[0074] In another embodiment, the first image in step S1 can be a bombyx mori image, when the first image is a bombyx mori image, the evaluation of the overall period of bombyx mori in the bombyx mori image is specifically:

[0075] Extract the bombyx mori characteristics of each bombyx mori in the bombyx mori image, including morphological characteristics, texture characteristics and color characteristics, and identify the period of each bombyx mori through a pre-trained second deep learning model according to the bombyx mori characteristics, if the number of larva period bombyx mori in the bombyx mori image is greater than the number of cocoon period bombyx mori, the bombyx mori image is a larva period image, otherwise it is a cocoon period image, if the number of larva period images is greater than the number of cocoon period images, the evaluation result of the overall period of bombyx mori is larva period, otherwise it is cocoon period.

[0076] The second deep learning model completes classification and recognition training by using a large number of larva period bombyx mori images and cocoon period bombyx mori images as training samples, and can identify and distinguish larva period bombyx mori and cocoon period bombyx mori in the image.

[0077] The larva period bombyx mori and the cocoon period bombyx mori usually have differences in morphology, texture and color, such as the larva period bombyx mori in long cylindrical shape, clear segment, clear segment texture, different body color according to the bombyx mori age, the cocoon period bombyx mori in the process of silk spinning, usually appears body type shortening, segment contraction and chest and abdomen yellow translucent, after completing the silk spinning cocoon, the cocoon shape is usually elliptical or beam waist type, the surface color is mainly white, yellow or green, etc., after the bombyx mori pupates in the cocoon, the pupa is initially light yellow, gradually changes to yellow or yellow brown with time, and finally usually turns to dark brown, therefore, the present application can also select the bombyx mori characteristics in the image to evaluate the overall period of bombyx mori in the silkworm room in this embodiment.

[0078] In step S1, the evaluation of the abnormal state of mulberry leaves through the mulberry leaf image is specifically:

[0079] S11, the mulberry leaf region is segmented in the mulberry leaf image, and the edge of each mulberry leaf is identified in the mulberry leaf region to obtain a plurality of mulberry sub-regions representing each mulberry leaf;

[0080] S12, the pixel values of each pixel point in each color channel of RGB in each mulberry sub-region are obtained respectively, and the channel mean value of each color channel corresponding to each mulberry sub-region is calculated respectively, and for each color channel mean value, it is judged whether it is outside the corresponding preset normal value range, if yes, the channel mean value is an abnormal channel mean value;

[0081] S13, the mulberry sub-region with abnormal channel mean value is regarded as the first abnormal region, if the ratio of the number of first abnormal regions to the number of mulberry sub-regions is less than a preset ratio, the mulberry leaf state is normal, otherwise the following steps are continued to be executed;

[0082] S14, obtain abnormal pixel points in each first abnormal region respectively, the abnormal pixel points are pixel points with one or more color channels whose pixel value difference with the abnormal channel mean value of the corresponding color channel in the first abnormal region is less than or equal to a preset difference value;

[0083] S15, in each first abnormal region, take the continuously adjacent abnormal pixel points as the second abnormal region, and obtain the shape features of each second abnormal region, and estimate the disease abnormality of the mulberry leaf through the third pre-trained deep learning model according to the shape features of each second abnormal region in each mulberry leaf region and the abnormal channel mean value of the existing color channel.

[0084] In step S11, the mulberry leaf region can be segmented and the edge of each mulberry leaf can be recognized through the existing threshold segmentation algorithm and edge detection algorithm, and a plurality of mulberry leaf regions representing each mulberry leaf are obtained, which is a specific implementation of the prior art and will not be described here.

[0085] In the above steps, the pixel values of the RGB three color channels are respectively the red channel pixel value, the green channel pixel value and the blue channel pixel value, which are usually represented by 8-bit unsigned integers to represent single-channel intensity, with a range of 0-255. The color of each pixel is composed of the values of the three channels, and the color visible to the human eye is simulated by different proportions of superposition. The channel mean value of the mulberry leaf region is the average of the sum of the pixel values of the corresponding color channel of all pixel points in the mulberry leaf region. The abnormality of mulberry leaf disease is usually accompanied by color change, so it can be judged by whether the channel mean value is abnormal. For example, the diseased area of the leaf (such as necrotic spots) will reflect more red light, the red channel pixel value will increase significantly, the red channel mean value will be higher than the normal red channel pixel value, the green channel pixel value will decrease when the chlorophyll is lost, the green channel mean value will be lower than the normal green channel pixel value, and so on.

[0086] In step S14, the abnormal pixel points are pixel points with one or more color channels whose pixel value difference with the abnormal channel mean value of the corresponding color channel in the first abnormal region is less than or equal to a preset difference value. For example, in a first abnormal region, there is an abnormal channel mean value in the red channel. If the absolute value of the difference between the pixel value of pixel point A in the first abnormal region and the abnormal channel mean value of the red channel in the first abnormal region is less than or equal to the preset difference value, then pixel point A is the abnormal pixel point of the first abnormal region.

[0087] In the mulberry leaf disease abnormality, some abnormalities can be evaluated only by color, and some diseases are accompanied by special shapes, such as some disease spots. Therefore, in order to more accurately evaluate the disease condition, the application introduces the shape features of the region formed by the continuously adjacent abnormal pixel points to evaluate the disease abnormality.

[0088] In step S1, the types of environmental parameters include temperature, humidity, and light intensity, etc.

[0089] In some embodiments, in the process of obtaining multiple environmental parameters of the silkworm room by the sensor module in step S1, each sensor is further temperature-compensated and corrected by a preset neural network model, and the hyperparameters of the neural network model are optimized by a white whale optimization algorithm in the training process.

[0090] The hyperparameters of the neural network model are optimized by a white whale optimization algorithm in the training process, specifically:

[0091] Sy1, initialize the hyperparameter space, randomly generate a preset number of white whales according to the hyperparameter space to form an initial population, and the hyperparameter space represents the type of hyperparameters to be optimized and the corresponding value optimization interval, and each white whale represents a combination of hyperparameters;

[0092] Sy2, for each white whale, train the neural network model to perform temperature compensation tasks, and calculate the fitness according to the respective training output results, which can be selected as mean square error or mean absolute error;

[0093] Sy3, select the individual with the highest fitness in the current population as the leader white whale;

[0094] Sy4, determine the position update step of the current iteration according to the balance factor of the current iteration, if the balance factor is greater than 0.5, the white whale position is updated by the exploration step of the white whale optimization algorithm, if the balance factor is less than or equal to 0.5, the white whale position is updated by the development step of the white whale optimization algorithm;

[0095] Sy5, determine whether the balance factor of the current iteration is less than the falling probability of the current iteration:

[0096] If not, update the white whale position with step Sy4, repeat steps Sy2 to Sy5 until the iteration reaches the preset maximum iteration value, and stop, and the value of the hyperparameters represented by the real-time leader white whale is used as the optimal solution;

[0097] If yes, further update the white whale position by the white whale position obtained by step Sy4 in the current iteration, and iterate, repeat steps Sy2 to Sy5 until the iteration reaches the preset maximum iteration value, and stop, and the value of the hyperparameters represented by the real-time leader white whale is used as the optimal solution.

[0098] In the above beluga whale optimization algorithm, the specific implementation of updating the beluga whale position through its exploration step, development step and whale falling step belongs to the prior art, and will not be described here.

[0099] In step S22, the evaluation of the overall age of the silkworm is specifically:

[0100] The color features, texture features and body length features of the corresponding larval stage silkworm individuals are extracted from each second silkworm image, and the age of the corresponding larval stage silkworm individuals is analyzed based on the features, and finally the age with the largest number of larval stage silkworm individuals is taken as the overall age of the silkworm.

[0101] The analysis of the age of the corresponding larval stage silkworm individuals can be estimated and analyzed by a preset fourth deep learning model.

[0102] During the larval stage of the silkworm, it is divided into one to five age stages, and in each age stage, the body color, pattern and body length of the silkworm usually change obviously, for example, in the first age stage, the newly hatched silkworm has a dark body color and a small body size; in the fourth age stage, the body color of the silkworm is usually deep, mostly gray or greenish gray, the body length can reach about 3-4 cm, and the body is relatively long; and in the fifth age stage, the body color of the silkworm gradually becomes lighter, usually greenish white, the body surface is smooth, the body size is plump, and the body length can reach about 6-8 cm. According to the silkworm variety, there are different patterns in color or shape, such as plain silkworms and surface patterns of black and tiger patterns, most of which are plain silkworms or patterns, all of the above are for illustration, and the user can set and train the above models according to the actual breeding situation.

[0103] In step S23, the overall age of the silkworm and the abnormal state of the mulberry leaf disease are combined to adjust the environmental parameters, specifically:

[0104] According to the overall age of the silkworm, a first adaptive range of each environmental parameter is obtained;

[0105] According to the abnormal state of the mulberry leaf disease, a second adaptive range of each environmental parameter is obtained;

[0106] If the first adaptive range and the second adaptive range of the corresponding environmental parameter have an intersection, the intersection is taken as a third adaptive range of the corresponding environmental parameter, and the corresponding environmental parameter is adjusted to be within the corresponding third adaptive range;

[0107] If the first adaptive range and the second adaptive range of the corresponding environmental parameter have no intersection, a preset weight of the overall age of the silkworm and a preset weight of the abnormal state of the mulberry leaf disease are obtained, if the preset weight of the overall age of the silkworm is greater than the preset weight of the abnormal state of the mulberry leaf disease, the corresponding environmental parameter is adjusted to be within the corresponding first adaptive range, otherwise it is adjusted to be within the corresponding second adaptive range.

[0108] The preset weight of the whole age period of the silkworm and the preset weight of the abnormal state of the mulberry leaf disease, i.e., a person sets the weight for each age period and each abnormal type of disease according to the demand and experience.

[0109] In step S32, the whole abnormal state of the silkworm is evaluated according to each third silkworm image.

[0110] The color feature, the shape feature and the cocoon silk color feature of the silkworm in the third silkworm image are extracted, and the first abnormal type existing in the corresponding cocooning period silkworm individual is analyzed;

[0111] In all third silkworm images, the number of occurrences of each first abnormal type is counted, and if the ratio of the number of occurrences of the first abnormal type to the number of third silkworm images is greater than or equal to a preset number ratio, the corresponding first abnormal type is a second abnormal type, and the set of each second abnormal type is taken as the whole abnormal state of the silkworm.

[0112] The first abnormal type existing in the corresponding cocooning period silkworm individual is analyzed, which can be evaluated and analyzed by the fifth deep learning model pre-trained according to the color feature, the shape feature and the cocoon silk color feature of the silkworm.

[0113] In step S33, the environmental parameters are adjusted according to the abnormal state of the silkworm, which can be adjusted by the staff according to the experience, or the optimization value of each environmental parameter can be analyzed by the sixth deep learning model pre-trained according to the whole abnormal state of the silkworm, and then each environmental parameter is adjusted to the corresponding optimization value.

[0114] Embodiment two

[0115] The application also provides a silkworm cultivation management system based on multi-source data, which specifically comprises:

[0116] The sensor module acquires a plurality of environmental parameters in the silkworm room in real time.

[0117] The first evaluation module evaluates the whole period of the silkworm through a plurality of first images, wherein the whole period of the silkworm includes the larva period and the cocooning period.

[0118] The second evaluation module randomly acquires a plurality of mulberry leaf images, and evaluates the abnormal state of the mulberry leaf disease through the mulberry leaf images.

[0119] The first adjustment module randomly acquires a plurality of second silkworm images of the larva period silkworm individual when the whole period of the silkworm is the larva period, evaluates the whole age period of the silkworm according to the second silkworm image, and adjusts the environmental parameters in combination with the whole age period of the silkworm and the abnormal state of the mulberry leaf disease.

[0120] The second adjusting module, when the whole silkworm is in cocooning period, randomly acquires a plurality of third silkworm images of silkworm individuals in cocooning period, evaluates the abnormal state of the whole silkworm according to each third silkworm image, and adjusts the environmental parameters according to the abnormal state of the silkworm.

[0121] Embodiment three

[0122] The application further provides an electronic device, comprising a processor, a sending device, an input device, an output device and a memory, the processor can be realized by a general CPU (Central Processing Unit), a microprocessor, an application specific integrated circuit or one or more integrated circuits, and is used for executing a related program to realize the technical solution provided by the embodiment of the application, the memory can be realized by a ROM (ReadOnly Memory), a static storage device, a dynamic storage device or a RAM (Random Access Memory), and is used for storing a computer program code, the computer program code comprises computer instructions, and when the processor executes the computer instructions, the electronic device executes the method in any one of the possible implementation manners described above.

[0123] Embodiment four

[0124] The application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program comprises program instructions, and the program instructions make the processor execute the method in any one of the possible implementation manners described above when the processor of the electronic device executes the program instructions.

[0125] The application has the following beneficial effects:

[0126] The application evaluates the abnormal state of mulberry leaf diseases through a mulberry leaf image, adjusts environmental parameters in the larva period in combination with the whole silkworm age and the abnormal state of mulberry leaf diseases, adjusts the environmental parameters according to the abnormal state of silkworm in the cocooning period, adjusts the environmental parameters in different cases, takes the different growth periods of silkworm as a horizontal dimension and introduces the abnormal state of mulberry leaf diseases as a vertical dimension, adjusts the environmental parameters in combination with multiple dimensions, improves the refinement and precision of silkworm cultivation and management, and has the advantages that:

[0127] The mulberry leaf region is segmented and the single leaf edge is recognized, the interference of a complex background on disease identification is effectively excluded, the RGB three-channel mean is used to comprehensively analyze the multidimensional information of a color space, the local continuous region analysis of abnormal pixels is combined, single pixel noise interference is avoided, robustness is enhanced, and the accuracy of disease abnormal analysis is improved.

[0128] In the description, reference to "one embodiment," "an example," "certain examples," etc., mean that a particular feature, structure, material, or characteristic being referred to is included in at least one embodiment or example of the disclosure. The appearances of the phrases "in one embodiment," "an example," "in certain examples," etc., in various places in the specification are not necessarily referring to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0129] In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware, or in the form of a software functional unit. When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially, or the part that contributes to the prior art, or all or a part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions used to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various other media that can store programs.

[0130] The above description is merely one specific implementation of the present application, which enables those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A silkworm breeding management method based on multi-source data, characterized by, The method comprises the following steps: Real-time acquisition of multiple environmental parameters in the silkworm room through a sensor module, evaluation of the overall period of the silkworm through multiple first images, the overall period of the silkworm including the larva period and the cocooning period; random acquisition of multiple mulberry leaf images, evaluation of the abnormal state of mulberry leaf diseases through the mulberry leaf images; If the overall period of the silkworm is the larva period, the following steps are executed: Random acquisition of multiple second silkworm images of silkworm individuals in the larva period; Evaluation of the overall age of the silkworm according to the second silkworm images; Adjustment of the environmental parameters in combination with the overall age of the silkworm and the abnormal state of mulberry leaf diseases; If the overall period of the silkworm is the cocooning period, the following steps are executed: Random acquisition of multiple third silkworm images of silkworm individuals in the cocooning period; Evaluation of the overall abnormal state of the silkworm according to each third silkworm image; Adjustment of the environmental parameters according to the abnormal state of the silkworm.

2. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, When the first image is a silkworm room image, the evaluation of the overall period of the silkworm through multiple first images is specifically: Through a pre-trained first deep learning model, it is identified whether there is a cocooning device in the silkworm room image, the silkworm room image with the cocooning device is taken as a first silkworm room image, and if the ratio of the number of first silkworm room images to the number of silkworm room images is greater than a first preset threshold, the evaluation result of the overall period of the silkworm is the cocooning period, otherwise it is the larva period.

3. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, When the first image is a silkworm image, the evaluation of the overall period of the silkworm through multiple first images is specifically: Extracting silkworm features of each silkworm in the silkworm image, including morphological features, texture features and color features, identifying the period of each silkworm through a pre-trained second deep learning model according to the silkworm features, if the number of silkworms in the larva period is greater than the number of silkworms in the cocooning period in the silkworm image, the silkworm image is a larva period image, otherwise it is a cocooning period image, if the number of larva period images is greater than the number of cocooning period images, the evaluation result of the overall period of the silkworm is the larva period, otherwise it is the cocooning period.

4. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, The evaluation of the abnormal state of mulberry leaf diseases through the mulberry leaf images is specifically: Segmenting the mulberry leaf region in the mulberry leaf image and identifying the edge of each mulberry leaf in the mulberry leaf region to obtain multiple mulberry sub-regions representing each mulberry leaf; Respectively acquiring the pixel values of each pixel in each color channel of RGB in each mulberry sub-region, and respectively calculating the channel mean value of each color channel corresponding to each mulberry sub-region, and for each color channel mean value, it is judged whether it is outside the corresponding preset normal value range, if so, the channel mean value is an abnormal channel mean value; The mulberry sub-region with the abnormal channel mean value is regarded as a first abnormal region, if the ratio of the number of first abnormal regions to the number of mulberry sub-regions is less than a preset ratio, the mulberry state is normal, otherwise the following steps are continued; Respectively acquiring abnormal pixel points in each first abnormal region, the abnormal pixel point being a pixel point with an absolute value of the difference between the pixel value of one or more color channels and the abnormal channel mean value of the corresponding color channel in the first abnormal region being less than or equal to a preset difference value; In each first abnormal region, continuously adjacent abnormal pixel points are taken as second abnormal regions, so as to obtain the shape features of each second abnormal region, and the third deep learning model is used to estimate the disease abnormality of the mulberry leaves according to the shape features of each second abnormal region in each mulberry leaf region and the abnormal channel mean value of the existing color channel.

5. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, In the process of acquiring the multiple environment parameters of the cocoon room by the sensor module, the sensor is also subjected to temperature compensation correction by a preset neural network model, and the neural network model is optimized by a white whale optimization algorithm in the training process.

6. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, The evaluation of the overall age of the domestic silkworm is specifically: Color features, texture features and body length features of the corresponding larval stage domestic silkworm individuals are extracted from each second domestic silkworm image, and the age of the corresponding larval stage domestic silkworm individuals is analyzed, and finally the age stage with the largest number of larval stage domestic silkworm individuals is taken as the overall age of the domestic silkworm.

7. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, The adjustment of the environment parameters in combination with the overall age of the domestic silkworm and the mulberry disease abnormality state is specifically: A first adaptive range of each environment parameter is obtained according to the overall age of the domestic silkworm; A second adaptive range of each environment parameter is obtained according to the mulberry disease abnormality state; If there is an intersection between the first adaptive range and the second adaptive range of the corresponding environment parameter, the intersection is taken as a third adaptive range of the corresponding environment parameter, and the corresponding environment parameter is adjusted to be within the corresponding third adaptive range; If there is no intersection between the first adaptive range and the second adaptive range of the corresponding environment parameter, preset weights of the overall age of the domestic silkworm and preset weights of the mulberry disease abnormality state are obtained, if the preset weight of the overall age of the domestic silkworm is greater than the preset weight of the mulberry disease abnormality state, the corresponding environment parameter is adjusted to be within the corresponding first adaptive range, otherwise, the corresponding environment parameter is adjusted to be within the corresponding second adaptive range.

8. The silkworm breeding management method based on multi-source data according to claim 1, characterized by, The evaluation of the overall abnormal state of the domestic silkworm according to each third domestic silkworm image is: Color features, morphological features and cocoon silk color features of the domestic silkworm are extracted from the third domestic silkworm image, and a first abnormal type existing in the corresponding cocooning period domestic silkworm individual is analyzed; In all third domestic silkworm images, the number of occurrences of each first abnormal type is counted, if the ratio of the number of occurrences of the first abnormal type to the number of third domestic silkworm images is greater than or equal to a preset number ratio, the corresponding first abnormal type is a second abnormal type, and a set of each second abnormal type is taken as the overall abnormal state of the domestic silkworm.

9. A silkworm breeding management system based on multi-source data, applied to the silkworm breeding management method based on multi-source data according to any one of claims 1 to 8, characterized in that, It comprises: A sensor module acquires multiple environment parameters in a cocoon room in real time; A first evaluation module evaluates the overall period of the domestic silkworm through multiple first images, and the overall period of the domestic silkworm includes a larval stage and a cocooning period; A second evaluation module randomly acquires multiple mulberry leaf images, and evaluates a mulberry disease abnormality state through the mulberry leaf images; A first adjustment module, when the overall period of the domestic silkworm is the larval stage, randomly acquires second domestic silkworm images of multiple larval stage domestic silkworm individuals, evaluates the overall age of the domestic silkworm according to the second domestic silkworm images, and adjusts the environment parameters in combination with the overall age of the domestic silkworm and the mulberry disease abnormality state. The second adjusting module, when the period in which the whole silkworm is located is cocooning period, randomly acquires third silkworm images of multiple silkworm individuals in cocooning period, evaluates abnormal state of the whole silkworm according to each third silkworm image, and adjusts environmental parameters according to the abnormal state of the silkworm.

10. The silkworm breeding management system based on multi-source data according to claim 9, characterized by, The silkworm leaf image is used to evaluate the abnormal state of the silkworm leaf disease, specifically: The silkworm leaf region is segmented in the silkworm leaf image, and the edge of each silkworm leaf is identified in the silkworm leaf region to obtain multiple silkworm sub-regions representing each silkworm leaf; The pixel values of each pixel in each silkworm sub-region in the RGB three color channels are acquired respectively, so as to calculate the channel mean value of each color channel corresponding to each silkworm sub-region, and for each color channel mean value, it is judged whether it is outside the corresponding preset normal value range, if yes, the channel mean value is an abnormal channel mean value; The silkworm sub-region with abnormal channel mean value is regarded as a first abnormal region, if the ratio of the number of first abnormal regions to the number of silkworm sub-regions is less than a preset ratio, the silkworm leaf state is evaluated as normal, otherwise the following steps are continued to be executed; Abnormal pixel points are acquired in each first abnormal region, the abnormal pixel points are pixel points with an absolute value of the difference between the pixel value of one or more color channels and the abnormal channel mean value of the corresponding color channel in the first abnormal region being less than or equal to a preset difference value; In each first abnormal region, continuously adjacent abnormal pixel points are taken as second abnormal regions, so as to obtain the shape features of each second abnormal region, and through the third deep learning model pre-trained, the abnormal state of the disease existing in the silkworm leaf is estimated according to the shape features of each second abnormal region in each silkworm sub-region and the abnormal channel mean value of the existing color channel.

Images