Intelligent breeding situation monitoring method, system, equipment and storage medium

Abstract

The present application relates to the field of data processing, and in particular to an intelligent farming monitoring method, system, device and storage medium, wherein the method comprises: acquiring farming images at first time intervals and image processing the farming images to obtain image processing results, the image processing results comprising length and width data and position coordinates of the farmed organisms in the farming images; calculating and generating the estimated number of farmed organisms in the farming images based on the image processing results and a preset assumed distribution model; acquiring and processing the image processing results to obtain the pile degree data of the farmed organisms in the farming images; acquiring and processing the image processing results to obtain the weight data of the farmed organisms in the farming images. The present application has the effect of facilitating the livestock farming conditions obtained through monitoring to be close to the actual livestock farming conditions.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an intelligent breeding situation monitoring method, system, equipment and storage medium. Background Art

[0002] With the rapid development of information technology, information technologies such as remote video monitoring and image processing have been widely used in the technical field of livestock breeding.

[0003] At present, livestock are monitored by video or picture through cameras in livestock farms, and the videos or pictures taken by the cameras are further acquired and processed by computers to obtain processing results, and the livestock breeding conditions in the livestock farms are understood through the processing results.

[0004] In the process of implementing the present application, the inventors found that the above-mentioned technology has at least the following problems: objective factors such as the light intensity in the livestock farm and the fact that some livestock are not photographed will cause the livestock breeding conditions obtained by the existing technology to deviate greatly from the actual livestock breeding conditions. Summary of the invention

[0005] In order to facilitate making the livestock breeding conditions obtained through monitoring close to the actual livestock breeding conditions, the present application provides an intelligent breeding condition monitoring method, system, equipment and storage medium.

[0006] In the first aspect, the present application provides an intelligent breeding situation monitoring method, which adopts the following technical solutions:

[0007] An intelligent breeding situation monitoring method, comprising:

[0008] Acquire a farming situation image once at a preset first time interval and process the farming situation image to obtain an image processing result, wherein the image processing result includes length and width data and position coordinates of the farmed organisms in the farming situation image;

[0009] Calculating a first position probability density data group corresponding to each of the farming situation images one by one according to the image processing result and a preset hypothetical distribution model;

[0010] Processing each of the first position probability density data groups to obtain a corresponding second position probability density data group;

[0011] Process each of the second position probability density data groups to generate a corresponding quantity estimation value, and process each of the quantity estimation values ​​to generate a first quantity array;

[0012] Processing the first number array to generate a second number array at every preset second time interval;

[0013] Processing the second quantity array to generate an estimated quantity of farmed organisms in the farmed condition image;

[0014] Acquiring and processing the image processing result to obtain data on the degree of clumping of the cultured organisms in the culture condition image;

[0015] The image processing result is acquired and processed to obtain the weight data of the farmed organisms in the farming situation image.

[0016] By adopting the above technical solution, after obtaining the length and width data and position coordinates of the farmed organisms, firstly, the dispersion degree of the farmed organisms on each farmed situation image is obtained through the position coordinates of the farmed organisms on the farmed situation image and the preset hypothetical distribution model, so as to obtain a farmed situation image with the best dispersion degree of the farmed organisms among all the farmed situation images, and then the number of farmed organisms in the farmed situation image is counted, and this number will be closer to the actual number of farmed organisms in the farm, so that it is convenient for users to remotely know the number of farmed organisms to determine whether the farmed organisms have been sold privately; secondly, the pile degree data of the farmed organisms can be calculated through the length and width data and position coordinates of the farmed organisms on each farmed situation image, so that it is convenient for users to determine whether the farmed organisms have temperature discomfort and other conditions based on the pile degree data; thirdly, the weight data of each farmed organism can be calculated through the length and width data of the farmed organisms on each farmed situation image, so that it is convenient for users to promptly know whether each farmed organism has reached the stage where it can be sold.

[0017] Preferably, the hypothetical distribution model is:

[0018] f(x,y) = 1 / (2*w+2*h-4)*C, (x or y = 0);

[0019] f(x,y) = 1 / [w*h - 2*(w+h-2) ]*(1 - C), (x or y≠0);

[0020] w is the length of the farming situation image, and h is the width of the farming situation image;

[0021] The first position probability density data set of each farmed organism in the farmed condition image calculated based on the image processing result and the preset hypothetical distribution model includes:

[0022] Acquire the position coordinates of each of the farmed organisms in the farming situation image;

[0023] The first position probability density data set is obtained by processing each of the position coordinates through the assumed distribution model.

[0024] By adopting the above technical solution, the location coordinates of the cultured organisms and the preset hypothetical distribution

[0025] The model can calculate the probability density of each farmed organism at the position coordinate. Since each livestock is located at a different position in the farm, each farmed organism has its corresponding probability density. The probability density corresponding to the farmed organisms on each farming situation image is combined into a first position probability density data group, which is convenient for subsequent comparison to understand the degree of dispersion of the farmed organisms on the farming situation image.

[0026] Preferably, the processing of each of the first position probability density data groups to obtain a corresponding second position probability density data group comprises:

[0027] Calculating the distance between each of the first position probability density data groups and a preset standard position probability density data group;

[0028] Each of the distances is compared with a preset first threshold, and the first position probability density data group whose distance is less than the first threshold is obtained and recorded as the second position probability density data group.

[0029] By adopting the above technical scheme, the distance obtained after comparing the first position probability density data group and the standard position probability density data group can reflect the gap between the dispersion degree of the farmed organisms in the farming situation image corresponding to the first position probability density data group and the ideal dispersion degree. The farming situation images with excessively large gaps are filtered out through the first threshold, because excessively large gaps also mean that the distribution of farmed organisms in the farming situation image is difficult to identify the number of farmed organisms more accurately.

[0030] Preferably, the processing of each of the second position probability density data groups to generate a corresponding quantity estimate, and the processing of each of the quantity estimates to generate a first quantity array comprises:

[0031] Counting the number of data in each of the second position probability density data groups, and recording the number of data as the number estimation value;

[0032] Each of the quantity estimates is combined to generate the first quantity array.

[0033] By adopting the above technical solution, the number of data in the second position probability density data group represents the estimated number of farmed organisms in each farming situation image, and the number estimation values ​​are organized into a first number array for comparison, so as to facilitate obtaining the number estimation value closest to the actual number of farmed organisms through the first skilled array.

[0034] Preferably, processing the first number array to generate the second number array at every preset second time interval comprises:

[0035] calculating a standard deviation of the quantity estimates in the first quantity array at every second time interval;

[0036] The quantity estimation values ​​exceeding a preset second threshold in the first quantity array are deleted to generate a second quantity array.

[0037] By adopting the above technical solution, some quantity estimation values ​​in the first quantity array are too different from the actual quantity of farmed organisms. According to historical data, when a quantity estimation value in the first quantity array exceeds the preset second threshold, it can be considered that such quantity estimation value is too different from the actual quantity of farmed organisms. Deleting such quantity estimation value can improve the effectiveness of the first quantity array.

[0038] Preferably, the processing of the second number array to generate an estimated number of farmed organisms in the farming situation image comprises:

[0039] Arrange the quantity estimation values ​​in the second quantity array to generate a third quantity array;

[0040] The estimated quantity is generated by performing mathematical statistics on the third quantity array.

[0041] By adopting the above technical solution, arranging the estimated quantity values ​​in the second quantity array helps to perform mathematical statistics on the estimated quantity values ​​in the third quantity array, so as to obtain the estimated quantity in the third quantity array that is closest to the actual number of farmed organisms.

[0042] Preferably, the acquiring and processing of the image processing result to obtain the data of the degree of clumping of the farmed organisms in the farming situation image comprises:

[0043] Processing the length and width data of each of the farmed organisms in the farmed condition image to generate the farthest distance of a cluster;

[0044] According to the farthest distance of the cluster and the position coordinates of each of the farmed organisms in the farming situation image, the image is divided into edge point coordinates and core point coordinates;

[0045] The stacking degree data is calculated according to the core point coordinates.

[0046] By adopting the above technical solution, the length and width data and position coordinates of the cultured organisms are used to determine the pile degree data of the cultured organisms, so that the user can judge whether the cultured organisms are in temperature discomfort or other conditions based on the pile degree data.

[0047] In the second aspect, the present application provides an intelligent breeding situation monitoring system, which adopts the following technical solutions:

[0048] An intelligent breeding situation monitoring system, comprising:

[0049] An image acquisition and processing module, used for acquiring a farming image once at a preset first time interval and image processing the farming image to obtain an image processing result, wherein the image processing result includes length and width data and position coordinates of the farmed organisms in the farming image;

[0050] A first position probability density data group generating module, used for calculating a first position probability density data group corresponding to each of the breeding situation images one by one according to the image processing result and a preset hypothetical distribution model;

[0051] A second position probability density data group generating module, used for processing each of the first position probability density data groups to obtain a corresponding second position probability density data group;

[0052] A first quantity array generating module, used for processing each of the second position probability density data groups to generate a corresponding quantity estimation value, and processing each of the quantity estimation values ​​to generate a first quantity array;

[0053] A second number array generating module, configured to process the first number array at every preset second time interval to generate a second number array;

[0054] an estimated number generating module, used for processing the second number array to generate an estimated number of farmed organisms in the farming situation image;

[0055] A clumping degree data generating module is used to obtain and process the image processing result to obtain the clumping degree data of the cultured organisms in the culture condition image;

[0056] The weight data generating module is used to obtain and process the image processing result to obtain the weight data of the farmed organism in the farming situation image.

[0057] By adopting the above technical solution, after the length and width data and position coordinates of the farmed organisms are obtained by the image acquisition and processing module, firstly, the dispersion degree of the farmed organisms on each farmed situation image is obtained by the position coordinates of the farmed organisms on the farmed situation image and the preset hypothetical distribution model, so as to obtain a farmed situation image with the best dispersion degree of the farmed organisms among all farmed situation images, and then the number of farmed organisms in the farmed situation image is counted, and this number will be closer to the actual number of farmed organisms in the farm, so that it is convenient for users to remotely know the number of farmed organisms to determine whether the farmed organisms have been sold privately; secondly, the length and width data and position coordinates of the farmed organisms on each farmed situation image are obtained by the pile degree data generation module, and then the pile degree data of the farmed organisms is calculated, so that it is convenient for users to judge whether the farmed organisms have temperature discomfort and other conditions according to the pile degree data; thirdly, the length and width data of the farmed organisms on each farmed situation image are obtained by the weight data generation module, and then the weight data of each farmed organism is calculated, so that it is convenient for users to timely know whether each farmed organism has reached the stage where it can be sold.

[0058] In a third aspect, the present application provides a computer device, which adopts the following technical solution: it includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute any of the above-mentioned intelligent breeding situation monitoring methods.

[0059] By adopting the above technical solution, after obtaining the length and width data and position coordinates of the farmed organisms, firstly, the dispersion degree of the farmed organisms on each farmed situation image is obtained through the position coordinates of the farmed organisms on the farmed situation image and the preset hypothetical distribution model, so as to obtain a farmed situation image with the best dispersion degree of the farmed organisms among all farmed situation images, and then count the number of farmed organisms in the farmed situation image, which will be closer to the actual number of farmed organisms in the farm, so that it is convenient for users to remotely know the number of farmed organisms to determine whether the farmed organisms have been sold privately; secondly, the pile degree data of the farmed organisms can be calculated through the length and width data and position coordinates of the farmed organisms on each farmed situation image, so that it is convenient for users to determine whether the farmed organisms are in temperature discomfort based on the pile degree data; thirdly, the weight data of each farmed organism can be calculated through the length and width data of the farmed organisms on each farmed situation image, so that it is convenient for users to promptly know whether each farmed organism has reached the stage where it can be sold.

[0060] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: storing a computer program that can be loaded by a processor and execute any of the above-mentioned intelligent breeding situation monitoring methods.

[0061] By adopting the above technical solution, after obtaining the length and width data and position coordinates of the farmed organisms, firstly, the dispersion degree of the farmed organisms on each farmed situation image is obtained through the position coordinates of the farmed organisms on the farmed situation image and the preset hypothetical distribution model, so as to obtain a farmed situation image with the best dispersion degree of the farmed organisms among all farmed situation images, and then count the number of farmed organisms in the farmed situation image, which will be closer to the actual number of farmed organisms in the farm, so that it is convenient for users to remotely know the number of farmed organisms to determine whether the farmed organisms have been sold privately; secondly, the pile degree data of the farmed organisms can be calculated through the length and width data and position coordinates of the farmed organisms on each farmed situation image, so that it is convenient for users to determine whether the farmed organisms are in temperature discomfort based on the pile degree data; thirdly, the weight data of each farmed organism can be calculated through the length and width data of the farmed organisms on each farmed situation image, so that it is convenient for users to promptly know whether each farmed organism has reached the stage where it can be sold.

[0062] In summary, the present application includes at least one of the following beneficial technical effects:

[0063] 1. After acquiring the breeding situation image, the length, width and position coordinates of the breeding organisms in the breeding situation image are obtained by processing the breeding situation image, and then the estimated number, pile degree data and weight data of the breeding organisms are calculated based on the length, width and position coordinates of the breeding organisms, so that the livestock breeding situation obtained through monitoring is close to the actual livestock breeding situation.

[0064] 2. By calculating the distance between the first position probability density data group and the standard position probability density data group, it is convenient to understand through the distance whether the breeding situation image corresponding to the first position probability density data group has reference value.

[0065] 3. When a quantity estimation value in the first quantity array exceeds a preset second threshold value, it can be considered that the difference between such quantity estimation value and the actual quantity of farmed organisms is too large. Deleting such quantity estimation value can improve the effectiveness of the first quantity array. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a flow chart of an intelligent breeding situation monitoring method in Example 1 of the present application.

[0067] Figure 2 It is a structural block diagram of an intelligent breeding situation monitoring system in Example 2 of the present application.

[0068] Explanation of the accompanying drawings: 100, image acquisition and processing module; 200, first position probability density data group generation module; 201, position coordinate acquisition module; 202, first position probability density data group calculation module; 300, second position probability density data group generation module; 301, distance calculation module; 302, second position probability density data group acquisition module; 400, first quantity array generation module; 401, quantity estimation value calculation module; 402, first quantity array acquisition module; 500, second quantity array generation module; 501, standard deviation calculation module; 502, second quantity array acquisition module; 600, estimated quantity generation module; 601, third quantity array generation module; 602, estimated quantity acquisition module; 700, stacking degree data generation module; 701, cluster maximum distance calculation module; 702, coordinate differentiation module; 703, stacking degree data acquisition module; 800, weight data generation module. DETAILED DESCRIPTION

[0069] The following is combined with Figure 1-2 This application is described in further detail.

[0070] Example 1

[0071] Example 1 of the present application discloses an intelligent breeding situation monitoring method. Figure 1 , methods include:

[0072] S100, acquiring a farming situation image once at a preset first time interval and image processing the farming situation image to obtain an image processing result, wherein the image processing result includes length and width data and position coordinates of the farmed organisms in the farming situation image.

[0073] In order to facilitate the visual understanding of the present technical solution, the farmed organisms are visualized as pigs in a pig farm.

[0074] The pig farm includes a number of pig pens, each of which houses a number of pigs, and a camera for photographing the pig pens is provided near each of the pig pens.

[0075] In implementation, taking one of the pig pens in a pig farm as an example, a video of the pig pen is captured in real time by a camera, and the captured video is transmitted to the server of the pig farm in real time. The server extracts a breeding situation image from the captured video at a first time interval, and then after the server receives each breeding situation image, the outline of each pig in the breeding situation image is first depicted by image processing, and then a rectangular frame that can frame the outline is made based on the outline, and then the rectangular frame corresponding to each pig is processed to obtain the length, width and position coordinates of each rectangular frame, the so-called length and width data including the length data and width data of the rectangular frame; after obtaining the pig pen image, a plane coordinate system about the pig pen is constructed based on the image of the pig pen, and the position coordinates are the coordinates of the center of the rectangular frame in the plane coordinate system.

[0076] S200, calculating and obtaining a first position probability density data group corresponding to each breeding situation image one by one according to the image processing result and a preset hypothetical distribution model.

[0077] S200 includes the following steps:

[0078] S201, obtaining the position coordinates of each farmed organism in the farming situation image.

[0079] Each identified pig in the pig pen image has its position coordinates. The server processes the pig pen image to obtain the two-dimensional position coordinates of each identified pig. Assuming that the number of identified pigs is n, the two-dimensional position coordinates of the pigs that can be obtained by the server are: (x1, y1), (x2, y2), ..., (xn, yn).

[0080] S202. Process each position coordinate by assuming a distribution model to obtain a first position probability density data set.

[0081] The two-dimensional probability density function of the location of the pigs in the assumed distribution model is summarized through historical data, and the two-dimensional probability density function is as follows:

[0082] f(x,y) = 1 / (2*w+2*h-4)*C, (x or y = 0);

[0083] f(x,y) = 1 / [w*h - 2*(w+h-2) ]*(1 - C), (x or y≠0);

[0084] It should be noted that x or y = 0 is used to represent that the pig is located at the edge of the pig pen, and x or y≠0 is used to represent that the pig is located in the middle of the pig pen; w is the length of the breeding situation image taken by the server, and h is the width of the breeding situation image taken by the server, and the length and width of each breeding situation image taken by the server are consistent; C in the above two-dimensional probability density function is a constant, and the constant is a constant summarized from historical data.

[0085] After the server obtains the position coordinates of each pig in a pig pen image, it calculates the two-dimensional probability density function value f(x, y) of each pig's position coordinates (x, y) through the above two-dimensional probability density function.

[0086] In implementation, if the server obtains the pig's two-dimensional position coordinates: (x1, y1),

[0087] (x2,y2),…,(xn,yn), then the calculated two-dimensional probability density function corresponds to f(x1,y1),f(x2,y2),…,f(xn,yn); and [f(x1,y1),f(x2,y2),…,f(xn,yn)] is recorded as the first position probability density data group.

[0088] S300, processing each first position probability density data group to obtain a corresponding second position probability density data group.

[0089] S300 includes the following steps:

[0090] S301, calculating the distance between each first position probability density data group and a preset standard position probability density data group.

[0091] The first position probability density data set in S202 can be understood as reflecting the dispersion of pigs in the pig pen image corresponding to the first position probability density data set. When the dispersion is relatively uniform, the number of pigs identified by the server is also relatively close to the actual number of pigs in the pig pen. The server obtains the corresponding probability density data set by simulating the situation when the pigs in the pig pen are uniformly dispersed and records it as the standard position probability density data set.

[0092] The server obtains the first position probability density data group and the preset standard position probability density data group, and further calculates the KL (Kullback-Leibler Divergence) distance between the first position probability density data group and the standard position probability density data group. The KL distance is used to characterize the degree of difference between the first position probability density data group and the standard position probability density data group.

[0093] S302, comparing each distance with a preset first threshold, obtaining a first position probability density data group whose distance is less than the first threshold and recording it as a second position probability density data group.

[0094] Each pigpen image has a corresponding first position probability density data group, but if the dispersion of pigs in a pigpen image is not good, then this pigpen image is not suitable for use as a reference image, and such an image needs to be screened out. The method for screening out such images in Example 1 of the present application is: first set a first threshold based on historical data, and calculate the KL distance between the first position probability density data group corresponding to each pigpen image and the standard position probability density data group. If the KL distance corresponding to an image is less than the first threshold, it means that the pigs in the image are relatively evenly dispersed, and the image has reference value, so the image is retained; if the KL distance corresponding to an image is not less than the first threshold, it means that the pigs in the image are unevenly dispersed, and the image has no reference value, so the image is removed; and the first position probability density data group with a KL distance less than the first threshold is recorded as the second position probability density data group.

[0095] S400, processing each second position probability density data group to generate a corresponding quantity estimation value, and processing each quantity estimation value to generate a first quantity array.

[0096] S400 includes the following steps:

[0097] S401. Count the number of data in each second position probability density data group, and record the number of data as a quantity estimation value.

[0098] Due to the different dispersion and positions of the pigs in each pig pen image, or the influence of factors such as the light intensity in the pig farm at that time, the number of pigs in each pig pen image identified by the server may be different, and thus the number of data in different first position probability density data groups may also be different, which in turn may cause the number of data in different second position probability density data groups to be different.

[0099] The server counts the number of data in each second position probability density data group, then obtains the number of data in each second position probability density data group, and records the number of data in the second position probability density data group as a quantity estimate, which is used to characterize the number of pigs identified in the pigpen image corresponding to the second position probability density data group.

[0100] S402: Combine each quantity estimation value to generate a first quantity array.

[0101] The quantity estimation value of each second position probability density data group is obtained through the server, and all the obtained quantity estimation values ​​are combined into a first quantity array.

[0102] For example, the server obtains m quantity estimation values ​​within a certain time range: A1, A2, ..., Am; further combine them into a first quantity array [A1, A2, ..., Am];

[0103] S500 , processing a first quantity array at every preset second time interval to generate a second quantity array.

[0104] S500 includes the following steps:

[0105] S501. Calculate the standard deviation of the quantity estimation values ​​in the first quantity array at every second time interval.

[0106] In one case, after the first second time interval, the first quantity array is: [A1, A2, ..., Am], the server obtains the data A1, A2, ..., Am in the first quantity array, and then calculates the standard deviation of A1, A2, ..., Am.

[0107] As time goes by, the number of pictures extracted by the server from the real-time video continues to increase. Accordingly, the number of first position probability density data groups and second position probability density data groups corresponding to the pictures will also increase. At the same time, the number of data corresponding to the second position probability density data group in the first quantity array will also increase.

[0108] For example, after the second second time interval, the first number array becomes [A1, A2, ..., Am, ..., Aa], where a is greater than m. Then, after the second second time interval, the standard deviation of the data set A1, A2, ..., Am, ..., Aa is calculated.

[0109] S502: Delete the quantity estimation values ​​exceeding a preset second threshold in the first quantity array to generate a second quantity array.

[0110] A second threshold is preset, and the second threshold is set to N times the standard deviation in combination with historical data, and N is set to 2 in Example 1 of the present application.

[0111] The data in the first quantity array are compared with the second threshold in sequence, the data in the first quantity array exceeding the second threshold is deleted through the server, and then the processed first quantity array is recorded as the second quantity array.

[0112] S600: Process the second quantity array to generate an estimated quantity of farmed organisms in the farming situation image.

[0113] S600 includes the following steps:

[0114] S601 , perform arrangement processing on the quantity estimation values ​​in the second quantity array to generate a third quantity array.

[0115] The server arranges the data in the second number array from small to large according to the magnitude of the values, and records the second number array after the arrangement as the third number array.

[0116] S602 , performing mathematical statistics on the third quantity array to generate an estimated quantity.

[0117] In Example 1 of the present application, mathematical statistics are performed on the third number array, that is, the median of the data in the third number array is selected, and the selected median is used as the estimated number of pigs in the pig pen.

[0118] The selected median is used as the estimated number of pigs in the pen because the median is not affected by the maximum and minimum extreme data in the third quantity array, and changes in some data have no effect on the median.

[0119] S700, obtaining and processing the image processing result to obtain the data of the pile-forming degree of the cultured organisms in the culture condition image.

[0120] S701, processing the length and width data of each farmed organism in the farming situation image to generate the farthest distance of the cluster.

[0121] During its operation, the server will obtain several images of farming conditions as time goes by. Taking one of the farming conditions images received by the server as an example, after the server processes the farming conditions image, it obtains the length data (w1, w1, ..., wb), width data (h1, h1, ..., hb) and position coordinates [(x1, y1), (x2, y2), ..., (xb, yb)] of each pig that can be identified in the farming conditions image. The premise of the above data is that the number of pigs that can be identified in the farming conditions image is b.

[0122] Furthermore, the average value w of the length data of each pig is calculated, and the average value h of the width data of each pig is also calculated, and then the obtained w and h are substituted into the calculation formula for calculating the farthest distance of the cluster:

[0123] The farthest distance of a cluster = sqrt((C1 * h)² + w²), where C1 is a constant summarized from historical data;

[0124] The cluster farthest distance can be obtained by substituting the obtained w and h into the calculation formula for calculating the cluster farthest distance.

[0125] S702. The position coordinates of each farmed organism in the farmed image are divided into edge point coordinates and core point coordinates according to the farthest distance of the cluster and the farmed image.

[0126] Several position coordinates can be identified on a breeding situation image. Taking one of the position coordinates as the base point, the base point is connected to all the position coordinates whose distance from the base point is not greater than the farthest distance of the cluster. Then, the position coordinates connected to the base point are used as the base point to connect all the position coordinates whose distance from the base point is not greater than the farthest distance of the cluster, and so on, until there are no more connectable position coordinates for the subsequent base points. These connected position coordinates form a cluster. In this way, several clusters can be obtained on a breeding situation image.

[0127] Through S701, the position coordinates of each pig in the breeding situation image [(x1, y1)、(x2, y2)、……、(xb, yb)] can be obtained, and a frame selection circle is further drawn with each position coordinate in each cluster as the center and the farthest distance of the cluster as the radius, that is, each position coordinate has its own frame selection circle, and the number of pig position coordinates in each frame selection circle is counted and recorded as the frame selection number. If the frame selection number in a frame selection circle is not greater than 1, the position coordinate corresponding to the frame selection circle is marked as the edge point coordinate; if the frame selection number in a frame selection circle is greater than 1, the position coordinate corresponding to the frame selection circle is marked as the core point coordinate.

[0128] S703, calculating the stacking degree data according to the core point coordinates.

[0129] Through S701 and S702, after obtaining a breeding situation image, the core point coordinates and edge coordinates of each cluster in the breeding situation image can be obtained. When calculating the pig pile degree data, it is necessary to count the number of core points in each cluster and obtain the corresponding number of core points, and then count the number of pig position coordinates in the breeding situation image and record it as the total number of pigs in a single pig pen. The calculation formula for the pig pile degree data in each cluster is as follows:

[0130] The degree of clustering data = core points / total number of pigs in a single pen;

[0131] Substituting the total number of pigs in a single pen and the number of core points in each cluster into the above-mentioned clustering degree data calculation formula, the clustering degree data of each cluster can be obtained.

[0132] S800, acquiring and processing the image processing result to obtain the weight data of the farmed organisms in the farming situation image.

[0133] Taking a breeding image obtained by the server as an example, the length and width data of the pigs identifiable in the breeding image are obtained through S100, and then the length and width data are substituted into the first function between the preset length and width data and the number of days for the pigs to grow, and the number of days for each identified pig to grow is further calculated.

[0134] Furthermore, the growth days of each pig are substituted into the second function between the preset growth days and the pig weight, and the weight data of each pig is further calculated.

[0135] The server will obtain the estimated number of pigs in the pig pen, the pigs' clustering degree data and the pigs' weight data and present them to the user through an interface display, so that the user can judge whether the pigs have been slaughtered privately based on the estimated number of pigs, whether the pigs are experiencing temperature discomfort based on the pigs' clustering degree data, and whether the corresponding pigs have reached the stage where they can be sold or whether the pigs have abnormal weight based on the pigs' weight data.

[0136] Figure 1 FIG. 1 is a flow chart of an intelligent breeding situation monitoring method in one embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but the steps are not necessarily executed in the order indicated by the arrows; unless otherwise specified in this document, there is no strict order restriction for the execution of the steps, and the steps may be executed in other orders; and Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0137] Example 2

[0138] Example 2 of the present application discloses an intelligent breeding situation monitoring system. Figure 2 , the intelligent breeding situation monitoring system includes:

[0139] The image acquisition and processing module 100 is used to acquire a farming situation image once at a preset first time interval and process the farming situation image to obtain an image processing result, wherein the image processing result includes the length and width data and position coordinates of the farmed organisms in the farming situation image.

[0140] The so-called length and width data include the length data and width data of the rectangular frame; after obtaining the pigpen image, a plane coordinate system about the pigpen is constructed based on the image of the pigpen, and the position coordinates are the coordinates of the center of the rectangular frame in the plane coordinate system.

[0141] The first position probability density data group generating module 200 is used to calculate the first position probability density data group corresponding to each breeding situation image according to the image processing result and the preset hypothetical distribution model.

[0142] The first position probability density data group generating module 200 includes:

[0143] The position coordinate acquisition module 201 is used to acquire the position coordinates of each farmed organism in the farming situation image.

[0144] The first position probability density data group calculation module 202 is used to process each position coordinate by assuming a distribution model to obtain a first position probability density data group.

[0145] The second position probability density data group generating module 300 is used to process each first position probability density data group to obtain a corresponding second position probability density data group.

[0146] The second position probability density data group generating module 300 includes: a distance calculation module 301, configured to calculate the distance between each first position probability density data group and a preset standard position probability density data group.

[0147] The second position probability density data group acquisition module 302 is used to compare each distance with a preset first threshold, obtain a first position probability density data group whose distance is less than the first threshold and record it as a second position probability density data group.

[0148] The first quantity array generating module 400 is used to process each second position probability density data group to generate a corresponding quantity estimation value, and process each quantity estimation value to generate a first quantity array.

[0149] The first quantity array generation module 400 includes:

[0150] The quantity estimation value calculation module 401 is used to count the number of data in each second position probability density data group and record the number of data as the quantity estimation value.

[0151] The first quantity array acquisition module 402 is used to combine each quantity estimation value to generate a first quantity array.

[0152] The second number array generating module 500 is configured to process the first number array at every preset second time interval to generate a second number array.

[0153] The second quantity array generating module 500 includes:

[0154] The standard deviation calculation module 501 is used to calculate the standard deviation of the quantity estimation value in the first quantity array at every second time interval.

[0155] The second quantity array acquisition module 502 is used to delete the quantity estimation values ​​exceeding a preset second threshold in the first quantity array to generate a second quantity array.

[0156] The estimated number generation module 600 is used to process the second number array to generate the estimated number of the farmed organisms in the farming situation image.

[0157] The estimated quantity generation module 600 includes:

[0158] The third quantity array generating module 601 is used to perform arrangement processing on the quantity estimation values ​​in the second quantity array to generate a third quantity array.

[0159] The estimated quantity acquisition module 602 is used to perform mathematical statistics on the third quantity array to generate an estimated quantity.

[0160] The clumping degree data generating module 700 is used to obtain and process the image processing result to obtain the clumping degree data of the farmed organisms in the farming situation image.

[0161] The cluster maximum distance calculation module 701 processes the length and width data of each farmed organism in the farming situation image to generate the cluster maximum distance.

[0162] The coordinate distinguishing module 702 is used to distinguish the position coordinates of each farmed organism in the farmed image into edge point coordinates and core point coordinates according to the farthest distance of the cluster and the farmed image.

[0163] The stacking degree data acquisition module 703 is used to calculate the stacking degree data according to the core point coordinates.

[0164] The weight data generating module 800 is used to obtain and process the image processing results to obtain the weight data of the farmed organisms in the farming situation image.

[0165] Example 3

[0166] In this embodiment 3, a computer device is disclosed, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned intelligent breeding condition monitoring method. The steps of the intelligent breeding condition monitoring method here can be the steps of the intelligent breeding condition monitoring method in each of the above-mentioned embodiments.

[0167] Example 4

[0168] In this embodiment 4, a computer-readable storage medium is disclosed, which stores a computer program that can be loaded by a processor and execute an intelligent breeding situation monitoring method as described above. The computer-readable storage medium includes, for example: 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 other media that can store program codes.

[0169] This specific embodiment is merely an explanation of the present invention and is not a limitation of the present invention. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed. However, as long as they are within the scope of the claims of the present invention, they are protected by the patent law.

Claims

1. An intelligent breeding situation monitoring method, characterized in that: include: Acquire a farming situation image once at a preset first time interval and process the farming situation image to obtain an image processing result, wherein the image processing result includes length and width data and position coordinates of the farmed organisms in the farming situation image; Calculating a first position probability density data group corresponding to each of the farming situation images one by one according to the image processing result and a preset hypothetical distribution model; Processing each of the first position probability density data groups to obtain a corresponding second position probability density data group; Process each of the second position probability density data groups to generate a corresponding quantity estimation value, and process each of the quantity estimation values ​​to generate a first quantity array; Processing the first number array to generate a second number array at every preset second time interval; Processing the second quantity array to generate an estimated quantity of farmed organisms in the farmed condition image; Acquiring and processing the image processing result to obtain data on the degree of clumping of the cultured organisms in the culture condition image; Acquiring and processing the image processing result to obtain weight data of the farmed organisms in the farming situation image; The assumed distribution model is: f(x,y) = 1 / (2*w+2*h-4)*C, (x or y = 0); f(x,y) = 1 / [w*h - 2*(w+h-2) ]*(1 - C), (x or y≠0); w is the length of the farming situation image, and h is the width of the farming situation image; The first position probability density data set of each farmed organism in the farmed condition image calculated based on the image processing result and the preset hypothetical distribution model includes: Acquire the position coordinates of each of the farmed organisms in the farming situation image; The first position probability density data set is obtained by processing each of the position coordinates through the assumed distribution model.

2. The intelligent breeding situation monitoring method according to claim 1 is characterized in that: The processing of each of the first position probability density data groups to obtain a corresponding second position probability density data group comprises: Calculating the distance between each of the first position probability density data groups and a preset standard position probability density data group; Each of the distances is compared with a preset first threshold, and the first position probability density data group whose distance is less than the first threshold is obtained and recorded as the second position probability density data group.

3. The intelligent breeding situation monitoring method according to claim 2 is characterized in that: The processing of each of the second position probability density data groups to generate a corresponding quantity estimation value, and the processing of each of the quantity estimation values ​​to generate a first quantity array comprises: Counting the number of data in each of the second position probability density data groups, and recording the number of data as the number estimation value; Each of the quantity estimates is combined to generate the first quantity array.

4. The intelligent breeding situation monitoring method according to claim 3 is characterized in that: The step of processing the first number array at every preset second time interval to generate a second number array comprises: calculating a standard deviation of the quantity estimates in the first quantity array at every second time interval; The quantity estimation values ​​exceeding a preset second threshold in the first quantity array are deleted to generate a second quantity array.

5. The intelligent breeding situation monitoring method according to claim 1 is characterized in that: The step of processing the second number array to generate an estimated number of farmed organisms in the farming situation image comprises: Arrange the quantity estimation values ​​in the second quantity array to generate a third quantity array; The estimated quantity is generated by performing mathematical statistics on the third quantity array.

6. The intelligent breeding situation monitoring method according to claim 1 is characterized in that: The step of acquiring and processing the image processing result to obtain the data of the degree of agglomeration of the farmed organisms in the farm condition image includes: Processing the length and width data of each of the farmed organisms in the farmed condition image to generate the farthest distance of a cluster; According to the farthest distance of the cluster and the position coordinates of each of the farmed organisms in the farming situation image, the position coordinates are divided into edge point coordinates and core point coordinates; The stacking degree data is calculated according to the core point coordinates.

7. An intelligent breeding situation monitoring system, characterized in that: include: An image acquisition and processing module (100) is used to acquire a farming situation image once at a preset first time interval and process the farming situation image to obtain an image processing result, wherein the image processing result includes length and width data and position coordinates of the farmed organisms in the farming situation image; A first position probability density data group generating module (200) is used to calculate, based on the image processing result and a preset hypothetical distribution model, a first position probability density data group corresponding to each of the breeding situation images; A second position probability density data group generating module (300), used for processing each of the first position probability density data groups to obtain a corresponding second position probability density data group; A first quantity array generating module (400) is used to process each of the second position probability density data groups to generate a corresponding quantity estimation value, and process each of the quantity estimation values ​​to generate a first quantity array; A second number array generating module (500), configured to process the first number array at every preset second time interval to generate a second number array; An estimated number generation module (600) is used to process the second number array to generate an estimated number of farmed organisms in the farming situation image; A clumping degree data generating module (700) is used to obtain and process the image processing result to obtain clumping degree data of the cultured organisms in the culture condition image; A weight data generating module (800) is used to obtain and process the image processing result to obtain the weight data of the farmed organism in the farming situation image; The assumed distribution model is: f(x,y) = 1 / (2*w+2*h-4)*C, (x or y = 0); f(x,y) = 1 / [w*h - 2*(w+h-2) ]*(1 - C), (x or y≠0); w is the length of the farming situation image, and h is the width of the farming situation image; The first position probability density data set of each farmed organism in the farmed condition image calculated based on the image processing result and the preset hypothetical distribution model includes: Acquire the position coordinates of each of the farmed organisms in the farming situation image; The first position probability density data set is obtained by processing each of the position coordinates through the assumed distribution model.

8. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executes any one of the intelligent breeding situation monitoring methods as claimed in claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and executes any one of the intelligent breeding situation monitoring methods according to claims 1 to 6.

Images