CN113068657A - Intelligent efficient pig raising method and system - Google Patents
Intelligent efficient pig raising method and system Download PDFInfo
- Publication number
- CN113068657A CN113068657A CN202110350581.1A CN202110350581A CN113068657A CN 113068657 A CN113068657 A CN 113068657A CN 202110350581 A CN202110350581 A CN 202110350581A CN 113068657 A CN113068657 A CN 113068657A
- Authority
- CN
- China
- Prior art keywords
- live pig
- live
- image
- pig
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K67/00—Rearing or breeding animals, not otherwise provided for; New breeds of animals
- A01K67/02—Breeding vertebrates
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K29/00—Other apparatus for animal husbandry
- A01K29/005—Monitoring or measuring activity, e.g. detecting heat or mating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
Abstract
The invention provides an intelligent high-efficiency pig raising method and system, wherein the method comprises the following steps: acquiring image information in each breeding fence, and separating a plurality of pieces of environment image information and live pig image information in the breeding fences according to the image information; identifying a live pig identification according to the live pig image, and acquiring a plurality of first images corresponding to the live pigs according to the live pig identification; acquiring an action track of each live pig corresponding to each live pig identification according to the plurality of first images; acquiring a second image according to the first image, inputting the second image into a live pig weight prediction model, and acquiring live pig weight information; recording the weight information of the live pigs according to a preset time interval, and acquiring the weight variation of the live pigs in the time interval; and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig. The invention can comprehensively analyze the growth condition of the live pig by combining environmental factors and a plurality of factors of the live pig, and realize the high-efficiency and accurate management of the live pig.
Description
Technical Field
The invention relates to the technical field of intelligent breeding, in particular to an intelligent efficient pig raising method and system.
Background
With the rapid development of domestic economy and the increase of income dominated by everyone, the consumption of meat by people is increased year by year, and the pig breeding faces the opportunity and simultaneously faces a great challenge. Traditional live pig breeding mainly adopts the family mode of breeding, but this kind of family mode of breeding's scale is little, the cultivation technique is laggard behind and is difficult to realize large-scale live pig and breeds, and the risk is great when the input time is long, is difficult to satisfy a large amount of live pig demands.
Therefore, the artificial intelligence technology begins to enter the field of live pig breeding, and can establish archives for each live pig through the intelligent ear tags, so that data of processes of breeding, quarantine, slaughtering, harmlessness and the like are recorded, data transparency in the live pig breeding process is realized, and the health condition of the live pigs is ensured. However, in the live pig breeding method in the prior art, the data recorded through artificial intelligence is single, and the judgment of the growth condition of the live pigs is not accurate enough.
Disclosure of Invention
Therefore, it is necessary to provide an intelligent and efficient pig raising method and system for solving the above technical problems.
An intelligent high-efficiency pig raising method comprises the following steps: acquiring image information in each breeding fence, and separating a plurality of environment image information and live pig image information in the breeding fence according to the image information, wherein the live pig image information comprises a plurality of live pig images, and the image information is acquired through camera equipment; identifying a live pig identification according to the live pig image, and acquiring a plurality of first images corresponding to the live pigs according to the live pig identification; acquiring an action track of each live pig corresponding to each live pig identification according to the plurality of first images; acquiring a second image according to the first image, inputting the second image into a live pig weight prediction model, and acquiring live pig weight information; recording the weight information of the live pigs according to a preset time interval, and acquiring the weight variation of the live pigs in the time interval; and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
In one embodiment, the identifying a live pig identifier according to the live pig image, and acquiring a plurality of first images corresponding to a live pig according to the live pig identifier specifically include: detecting at least one live pig mark according to the live pig image, wherein each live pig corresponds to a unique live pig mark; searching a corresponding first image in the live pig image information according to the detected live pig identification; and classifying the first images according to the live pig identifications to obtain a plurality of first images of the live pigs corresponding to each live pig identification.
In one embodiment, the obtaining of the action trajectory of each live pig corresponding to each live pig identifier according to the plurality of first images specifically includes: arranging the plurality of first images according to a time sequence; identifying the position of the live pig corresponding to the live pig identification in the first images, and confirming the position information of the live pig in the breeding fence; and sequentially connecting the position information according to the arrangement of the first images to obtain the action track of the live pig.
In one embodiment, the obtaining a second image according to the first image, inputting the second image into a live pig weight prediction model, and obtaining live pig weight information specifically includes: performing clear processing on the first image by adopting an image super-resolution reconstruction technology to obtain a second image; constructing an initial live pig weight prediction model according to a neural network algorithm; acquiring sample image information and sample weight information of a live pig to train the initial live pig weight prediction model to obtain a live pig weight prediction model; and inputting the second image into the live pig weight prediction model to obtain live pig weight information.
In one embodiment, the inputting the second image into the live pig weight prediction model to obtain live pig weight information specifically includes: the live pig weight prediction model comprises a first prediction network and a second prediction network; marking the contour of the live pig in the second image according to the first prediction network, and acquiring a contour image of the live pig corresponding to the live pig identification; acquiring feature information of the second image according to the contour image, wherein the feature information comprises one or more of coordinate information of the contour image, the number of pixel points corresponding to the contour image, and distance information and angle information between the contour image and the camera equipment; and inputting the characteristic information of the second image into the second prediction network to obtain the weight information of the live pig in the second image.
In one embodiment, after the weight variation of the live pig in the time interval is obtained according to the weight information of the live pig recorded in the preset time interval, before the determining the growth condition of the live pig according to the environment image information, the weight variation of the live pig and the action track, the method further includes: acquiring the temperature information of the live pigs in the breeding fence through infrared temperature measurement, and judging whether the temperature of the live pigs is normal or not; if the temperature information of the live pig is abnormal, marking the live pig identification corresponding to the live pig as an abnormal identification; and if the temperature information of the live pig is normal, marking the live pig identification corresponding to the live pig as a normal identification, and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
In one embodiment, if the temperature information of the live pig is normal, marking the live pig identifier corresponding to the live pig as a normal identifier, and determining the growth condition of the live pig according to the environment image information, the weight variation and the action trajectory of the live pig specifically includes: obtaining average variation according to the variation of the weight of all live pigs in the breeding fence; searching for the live pig with the weight variation smaller than the average variation, and marking the live pig identification corresponding to the live pig as the identification to be confirmed; acquiring the action track and the environment image information of the live pig according to the identification to be confirmed, and comprehensively judging the growth condition of the live pig according to the action track and the environment image information of the live pig; if the action track of the live pig and the environment image information of the live pig are normal, modifying the mark to be confirmed corresponding to the live pig as a normal mark; if the action track of the live pig and the environment image information of the live pig are abnormal, modifying the identification to be confirmed corresponding to the live pig as an abnormal identification; if the action track of the live pig is abnormal and the environmental image information is normal, modifying the mark to be confirmed corresponding to the live pig as a track abnormal mark; and if the action track of the live pig is normal and the environment image information is abnormal, modifying the identification to be confirmed corresponding to the live pig as an environment abnormal identification.
An intelligent high-efficiency pig raising system, comprising: the information separation module is used for acquiring the image information in each breeding fence and separating out first image information and environment image information of the pigs in the breeding fence according to the image information; the image acquisition module is used for identifying the live pig identification according to the first image information and acquiring a plurality of first images corresponding to the live pigs according to the live pig identification; the action track acquisition module is used for acquiring the action track of each live pig corresponding to each live pig identifier according to the plurality of first images; the weight information acquisition module is used for acquiring a second image according to the first image, inputting the second image into a live pig weight prediction model and acquiring live pig weight information; the weight variation acquiring module is used for recording the weight information of the live pigs in a plurality of time periods according to a preset time interval and acquiring the weight variation of the live pigs in adjacent time intervals; and the growth condition judging module is used for judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
Compared with the prior art, the invention has the advantages and beneficial effects that: the method can comprehensively analyze the growth condition of the live pigs by combining environmental factors and multiple factors of the live pigs, obtain more comprehensive and accurate analysis results, know the actual growth condition of the live pigs, and correspondingly process the live pigs with different growth conditions, thereby realizing the efficient and accurate management of the live pigs.
Drawings
FIG. 1 is a schematic flow chart of an intelligent high-efficiency pig raising method according to an embodiment;
fig. 2 is a schematic structural diagram of an intelligent efficient pig raising system in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings by way of specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In one embodiment, as shown in fig. 1, an intelligent and efficient pig raising method is provided, which comprises the following steps:
step S101, obtaining image information in each breeding fence, separating a plurality of environment image information and live pig image information in the breeding fence according to the image information, wherein the live pig image information comprises a plurality of live pig images, and the image information is obtained through a camera device.
Specifically, camera equipment is arranged at the same position above each breeding fence, image information in each breeding fence is obtained according to the camera equipment, and a plurality of pieces of environment image information and live pig image information in the breeding fence are separated according to the image information. Since a plurality of pigs are raised in the breeding fence, the pig image information includes a plurality of pig images.
And S102, identifying a live pig identification according to the live pig image, and acquiring a plurality of first images corresponding to the live pig according to the live pig identification.
Specifically, each live pig is worn with a corresponding ear tag, a live pig identification can be identified according to the ear tag of the live pig in the live pig image, and a plurality of first images of the live pig are obtained according to the live pig identification.
Specifically, the plurality of first images comprise an image of the live pig standing, and the plurality of first images of the live pig can be acquired when the live pig drinks or eats water, so that the first images can identify the overall condition of the live pig.
And S103, acquiring the action track of each live pig corresponding to each live pig identification according to the plurality of first images.
Specifically, a plurality of first images of the pig can be acquired in a time period, and the action track of the pig in the time period can be acquired by arranging the plurality of first images according to the time sequence.
And step S104, acquiring a second image according to the first image, inputting the second image into the live pig weight prediction model, and acquiring live pig weight information.
Specifically, the first image may have a problem of being not clear enough or other problems, so that the first image may be preprocessed to obtain a second image, the second image is input into a live pig weight prediction model, and live pig weight information is obtained by predicting according to the live pig weight prediction model.
And step S105, recording the weight information of the live pigs according to a preset time interval, and acquiring the weight variation of the live pigs in the time interval.
Specifically, a time interval is preset, live pig weight information is recorded according to the time interval, and the weight variation of a live pig in the time interval is obtained.
And step S106, judging the growth condition of the live pig according to the environment image information, the weight variation of the live pig and the action track of the live pig.
Specifically, the growth condition of the live pig is judged according to the environment image information, the live pig weight variation and the live pig action track.
In the embodiment, the image information in each breeding fence is obtained through the camera device, a plurality of environment image information and live pig image information in the breeding fence are separated according to the image information, a live pig identification is identified according to the live pig image information, a plurality of first images corresponding to the live pig are obtained according to the live pig identification, a motion track of the live pig corresponding to each live pig identification is obtained according to the first images, a second image is obtained according to the first images, the second image is input into the live pig weight prediction model to obtain the live pig weight information, the live pig weight information is recorded according to a preset time interval, the weight variation of the live pig in the time interval is obtained, the growth condition of the live pig is judged according to the environment image information, the weight variation of the live pig and the motion track, the growth condition of the live pig can be comprehensively analyzed by combining environmental factors and a plurality of factors of the live pig, and a more comprehensive and accurate analysis result is obtained, the method has the advantages that the actual growth conditions of the live pigs are known, and the live pigs with different growth conditions are correspondingly processed, so that the efficient and accurate management of the live pigs is realized.
Wherein, step S102 specifically includes: detecting at least one live pig mark according to the live pig image, wherein each live pig corresponds to a unique live pig mark; searching a corresponding first image in the live pig image information according to the detected live pig identification; and classifying the first images according to the live pig identifications to obtain a plurality of first images of the live pigs corresponding to each live pig identification.
Specifically, each live pig carries an intelligent ear tag, and the intelligent ear tag is provided with the only live pig identification of each live pig, and the live pig is identified according to the live pig identification. Detecting at least one live pig identification according to the first image information, searching a corresponding first image in the image information according to the detected live pig identification, classifying the first image according to the live pig identification, and acquiring a plurality of first images corresponding to each live pig, wherein the first images comprise images of each live pig.
Wherein, step S103 specifically includes: arranging a plurality of first images according to a time sequence; identifying the position of the live pig corresponding to the live pig identification in the first images, and confirming the position information of the live pig in the breeding fence; and connecting the position information in sequence according to the arrangement of the first images to obtain the action track of the live pig.
Specifically, the first images corresponding to the live pig identifications are arranged according to the time sequence, the positions of the live pigs corresponding to the live pig identifications in the first images are identified, the position information of the live pigs in the breeding fence is confirmed, the position information is connected once according to the arrangement of the first images, and the action tracks of the live pigs in the breeding fence in a time period are obtained.
Wherein, step S104 specifically includes: performing clear processing on the first image by adopting an image super-resolution reconstruction technology to obtain a second image; constructing an initial live pig weight prediction model according to a neural network algorithm; acquiring sample image information and sample weight information of a live pig to train an initial live pig weight prediction model to obtain a live pig weight prediction model; and inputting the second image into the live pig weight prediction model to obtain live pig weight information.
Specifically, because the image shot by the camera may have a problem of being not clear enough, the first image can be clearly processed by adopting an image super-resolution reconstruction technology to obtain a second image with higher resolution, so that the live pig identification in the second image can be conveniently identified. The first images can be images of the live pigs in states of drinking water or eating and the like in the plurality of first images, and more accurate weight information of the live pigs can be predicted through the standing images of the live pigs.
Specifically, an initial live pig weight prediction model is constructed according to a neural network algorithm, such as a Back Propagation (BP) neural network algorithm, sample image information and sample weight information of a live pig are collected, the initial live pig weight prediction model is trained, and the trained live pig weight prediction model is obtained; and inputting the second image into the live pig weight model so as to obtain the live pig weight information.
Inputting the second image into the live pig weight prediction model to obtain live pig weight information, wherein the method specifically comprises the following steps: the live pig weight prediction model comprises a first prediction network and a second prediction network; marking the contour of the live pig in the second image according to the first prediction network to obtain a contour image; acquiring feature information of a second image according to the contour image, wherein the feature information comprises one or more of coordinate information of the contour image, the number of pixel points corresponding to the contour image, and distance information and angle information between the contour image and the camera equipment; and inputting the characteristic information of the second image into a second prediction network to obtain the weight information of the live pig in the second image.
Specifically, the live pig weight prediction model can comprise a first prediction network and a second prediction network. The first prediction network is used for marking the live pig contour corresponding to the live pig identification and acquiring a contour image of the live pig; and acquiring the characteristic information of the second image according to the contour image, wherein the characteristic information can be one or more of coordinate information of the contour image in the second image, the number of pixel points corresponding to the contour image, and distance information and angle information between the contour image and the assumed device, so that the weight information of the live pig is predicted according to the characteristic information. And the second prediction network is used for acquiring the weight information of the live pig corresponding to the live pig mark through the analysis of the characteristic information of the second image.
After step S105 and before step S106, the method further includes: acquiring the temperature information of the live pigs in the breeding fence through infrared temperature measurement, and judging whether the temperature of the live pigs is normal or not; if the temperature information of the live pig is abnormal, marking the live pig identification corresponding to the live pig as an abnormal identification; and if the temperature information of the live pig is normal, marking the live pig identification corresponding to the live pig as a normal identification, and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
Specifically, the temperature information of the live pigs in the breeding fence is obtained through infrared temperature measuring equipment, whether the temperature of the live pigs in the breeding fence is normal or not is judged, and the growth conditions of the live pigs are judged according to the normality or not of the temperature information. If the temperature of the live pig is normal, the growth condition of the live pig is determined to be good; if the temperature of the live pig is abnormal, the growth condition of the live pig is determined to be poor, and the live pig with abnormal temperature is isolated and treated in time.
Specifically, the live pigs can be correspondingly marked according to the temperature information of the live pigs, the corresponding live pig identification is marked as an abnormal identification when the temperature information of the live pigs is abnormal, and the abnormal identification is sent to a breeder, so that the breeder can conveniently process the abnormal identification in time; and if the temperature information of the live pig is normal, marking the corresponding live pig identification as a normal identification, and further analyzing according to the weight variation of the live pig, the environmental image information and the action track.
When the temperature information of the live pig is normal, marking the live pig identification mark corresponding to the live pig as a normal identification, and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig, specifically comprising: obtaining average variation according to the variation of the weight of all live pigs in the breeding fence; searching for the live pig with the weight variation smaller than the average variation, and marking the live pig identification corresponding to the live pig as the identification to be confirmed; acquiring action tracks and environment image information of the pigs according to the identification to be confirmed, and comprehensively judging the growth conditions of the pigs according to the action tracks and the environment image information; if the action track of the live pig and the environment image information of the live pig are normal, modifying the mark to be confirmed corresponding to the live pig as a normal mark; if the action track of the live pig and the environment image information of the live pig are abnormal, modifying the identification to be confirmed corresponding to the live pig as an abnormal identification; if the action track of the live pig is abnormal and the environmental image information is normal, modifying the mark to be confirmed corresponding to the live pig as a track abnormal mark; and if the action track of the live pig is normal and the environment image information is abnormal, modifying the identification to be confirmed corresponding to the live pig as an environment abnormal identification.
Specifically, according to the weight variation of all live pigs in one breeding fence, the average variation of the live pigs in the breeding fence in one time period is obtained, the live pigs with the weight variation smaller than the average variation are screened out, and the live pig identification corresponding to the live pigs is marked as the identification to be confirmed.
Specifically, when the action track of the live pig and the environment image information of the live pig are normal, the live pig is determined to grow normally, and the corresponding mark to be confirmed is modified into a normal mark; when the action track of the live pig and the environment image information where the live pig is located are abnormal, possibly due to environment, the action track of the live pig is abnormal, the growth of the live pig is determined to be abnormal, and the corresponding mark to be confirmed is modified into an abnormal mark; when the action track of the live pig is abnormal and the image information of the environment is normal, the live pig may form an abnormal action track due to reasons other than the environment, so that a breeder can perform further investigation and modify the corresponding to-be-confirmed identifier into a track abnormal identifier; when the action track of the live pig is normal and the image information of the environment is abnormal, the abnormal situation is shown in the growth environment of the live pig, but the adverse effect on the growth of the live pig is not yet realized, so that the environment can be correspondingly processed, the subsequent adverse effect is avoided, and the corresponding mark to be confirmed is modified into the abnormal environment mark.
As shown in fig. 2, there is provided an intelligent high-efficiency pig raising system 20, comprising: the device comprises an information separation module 21, an image acquisition module 22, an action track acquisition module 23, a weight information acquisition module 24, a weight variation acquisition module 25 and a growth condition judgment module 26, wherein:
the information separation module 21 is configured to acquire image information in each breeding fence, and separate a plurality of pieces of environment image information and live pig image information in the breeding fence according to the image information, where the live pig image information includes a plurality of live pig images, and the image information is acquired by a camera device;
the image acquisition module 22 is used for identifying a live pig identifier according to the live pig image and acquiring a plurality of first images corresponding to the live pig according to the live pig identifier;
the action track obtaining module 23 is configured to obtain an action track of each live pig corresponding to each live pig identifier according to the plurality of first images;
the weight information acquisition module 24 is used for acquiring a second image according to the first image, inputting the second image into the live pig weight prediction model and acquiring live pig weight information;
the weight variation obtaining module 25 is configured to record weight information of the live pigs according to a preset time interval, and obtain weight variations of the live pigs within the time interval;
and the growth condition judging module 26 is used for judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
In one embodiment, the image acquisition module 22 is specifically configured to: detecting at least one live pig mark according to the live pig image, wherein each live pig corresponds to a unique live pig mark; searching a corresponding first image in the live pig image information according to the detected live pig identification; and classifying the first images according to the live pig identifications to obtain a plurality of first images of the live pigs corresponding to each live pig identification.
In one embodiment, the action trajectory acquiring module 23 is specifically configured to: arranging a plurality of first images according to a time sequence; identifying the positions of the live pigs corresponding to the live pig identifications in the first images, and confirming the position information of the live pigs in the breeding fence; and connecting the position information in sequence according to the arrangement of the first images to obtain the action track of the live pig.
In one embodiment, the weight information obtaining module 24 is specifically configured to: performing clear processing on the first image by adopting an image super-resolution reconstruction technology to obtain a second image; constructing an initial live pig weight prediction model according to a neural network algorithm; acquiring sample image information and sample weight information of a live pig to train an initial live pig weight prediction model to obtain a live pig weight prediction model; and inputting the second image into the live pig weight prediction model to obtain live pig weight information.
In one embodiment, the growth status determination module 24 is further configured to: acquiring the temperature information of the live pigs in the breeding fence through infrared temperature measurement, and judging whether the temperature of the live pigs is normal or not; if the temperature information of the live pig is abnormal, marking the live pig identification corresponding to the live pig as an abnormal identification; and if the temperature information of the live pig is normal, marking the live pig identification corresponding to the live pig as a normal identification, and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
In one embodiment, the growth status determination module 24 is further configured to: obtaining average variation according to the variation of the weight of all live pigs in the breeding fence; searching for the live pig with the weight variation smaller than the average variation, and marking the live pig identification corresponding to the live pig as the identification to be confirmed; acquiring action tracks and environment image information of the pigs according to the identification to be confirmed, and comprehensively judging the growth conditions of the pigs according to the action tracks and the environment image information; if the action track of the live pig and the environment image information of the live pig are normal, modifying the mark to be confirmed corresponding to the live pig as a normal mark; if the action track of the live pig and the environment image information of the live pig are abnormal, modifying the identification to be confirmed corresponding to the live pig as an abnormal identification; if the action track of the live pig is abnormal and the environmental image information is normal, modifying the mark to be confirmed corresponding to the live pig as a track abnormal mark; and if the action track of the live pig is normal and the environment image information is abnormal, modifying the identification to be confirmed corresponding to the live pig as an environment abnormal identification.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
It will be apparent to those skilled in the art that the modules or steps of the invention described above may be implemented in a general purpose computing device, they may be centralized on a single computing device or distributed across a network of computing devices, and optionally they may be implemented in program code executable by a computing device, such that they may be stored on a computer storage medium (ROM/RAM, magnetic disks, optical disks) and executed by a computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
The foregoing is a more detailed description of the present invention that is presented in conjunction with specific embodiments, and the practice of the invention is not to be considered limited to those descriptions. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (8)
1. An intelligent high-efficiency pig raising method is characterized by comprising the following steps:
acquiring image information in each breeding fence, and separating a plurality of environment image information and live pig image information in the breeding fence according to the image information, wherein the live pig image information comprises a plurality of live pig images, and the image information is acquired through camera equipment;
identifying a live pig identification according to the live pig image, and acquiring a plurality of first images corresponding to the live pigs according to the live pig identification;
acquiring an action track of each live pig corresponding to each live pig identification according to the plurality of first images;
acquiring a second image according to the first image, inputting the second image into a live pig weight prediction model, and acquiring live pig weight information;
recording the weight information of the live pigs according to a preset time interval, and acquiring the weight variation of the live pigs in the time interval;
and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
2. The intelligent and efficient pig raising method according to claim 1, wherein the identifying a pig id according to the pig image, and the obtaining a plurality of first images corresponding to a pig according to the pig id specifically comprises:
detecting at least one live pig mark according to the live pig image, wherein each live pig corresponds to a unique live pig mark;
searching a corresponding first image in the live pig image information according to the detected live pig identification;
and classifying the first images according to the live pig identifications to obtain a plurality of first images of the live pigs corresponding to each live pig identification.
3. The intelligent and efficient pig raising method according to claim 1, wherein the obtaining of the action trajectory of each pig, corresponding to each pig identifier, according to the plurality of first images specifically comprises:
arranging the plurality of first images according to a time sequence;
identifying the position of the live pig corresponding to the live pig identification in the first images, and confirming the position information of the live pig in the breeding fence;
and sequentially connecting the position information according to the arrangement of the first images to obtain the action track of the live pig.
4. The intelligent and efficient pig raising method according to claim 1, wherein the obtaining of the second image according to the first image and the inputting of the second image into a live pig weight prediction model to obtain the live pig weight information specifically comprises:
performing clear processing on the first image by adopting an image super-resolution reconstruction technology to obtain a second image;
constructing an initial live pig weight prediction model according to a neural network algorithm;
acquiring sample image information and sample weight information of a live pig to train the initial live pig weight prediction model to obtain a live pig weight prediction model;
and inputting the second image into the live pig weight prediction model to obtain live pig weight information.
5. The intelligent and efficient pig raising method according to claim 4, wherein the step of inputting the second image into the live pig weight prediction model to obtain the live pig weight information specifically comprises the steps of:
the live pig weight prediction model comprises a first prediction network and a second prediction network;
marking the contour of the live pig in the second image according to the first prediction network, and acquiring a contour image of the live pig corresponding to the live pig identification;
acquiring feature information of the second image according to the contour image, wherein the feature information comprises one or more of coordinate information of the contour image, the number of pixel points corresponding to the contour image, and distance information and angle information between the contour image and the camera equipment;
and inputting the characteristic information of the second image into the second prediction network to obtain the weight information of the live pig in the second image.
6. The intelligent and efficient pig raising method according to claim 1, further comprising, after recording weight information of a live pig according to a preset time interval and obtaining a weight variation of the live pig within the time interval, before judging the growth condition of the live pig according to the environment image information, the weight variation of the live pig and the action trajectory:
acquiring the temperature information of the live pigs in the breeding fence through infrared temperature measurement, and judging whether the temperature of the live pigs is normal or not;
if the temperature information of the live pig is abnormal, marking the live pig identification corresponding to the live pig as an abnormal identification;
and if the temperature information of the live pig is normal, marking the live pig identification corresponding to the live pig as a normal identification, and judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
7. The intelligent and efficient pig raising method according to claim 6, wherein if the temperature information of the live pig is normal, the live pig identifier corresponding to the live pig is marked as a normal identifier, and the growth condition of the live pig is judged according to the environment image information, the weight variation and the action trajectory of the live pig, which specifically comprises:
obtaining average variation according to the variation of the weight of all live pigs in the breeding fence;
searching for the live pig with the weight variation smaller than the average variation, and marking the live pig identification corresponding to the live pig as the identification to be confirmed;
acquiring the action track and the environment image information of the live pig according to the identification to be confirmed, and comprehensively judging the growth condition of the live pig according to the action track and the environment image information of the live pig;
if the action track of the live pig and the environment image information of the live pig are normal, modifying the mark to be confirmed corresponding to the live pig as a normal mark;
if the action track of the live pig and the environment image information of the live pig are abnormal, modifying the identification to be confirmed corresponding to the live pig as an abnormal identification;
if the action track of the live pig is abnormal and the environmental image information is normal, modifying the mark to be confirmed corresponding to the live pig as a track abnormal mark;
and if the action track of the live pig is normal and the environment image information is abnormal, modifying the identification to be confirmed corresponding to the live pig as an environment abnormal identification.
8. The utility model provides an intelligent high-efficient pig raising system which characterized in that includes:
the information separation module is used for acquiring the image information in each breeding fence and separating out first image information and environment image information of the pigs in the breeding fence according to the image information;
the image acquisition module is used for identifying the live pig identification according to the first image information and acquiring a plurality of first images corresponding to the live pigs according to the live pig identification;
the action track acquisition module is used for acquiring the action track of each live pig corresponding to each live pig identifier according to the plurality of first images;
the weight information acquisition module is used for acquiring a second image according to the first image, inputting the second image into a live pig weight prediction model and acquiring live pig weight information;
the weight variation acquiring module is used for recording the weight information of the live pigs in a plurality of time periods according to a preset time interval and acquiring the weight variation of the live pigs in adjacent time intervals;
and the growth condition judging module is used for judging the growth condition of the live pig according to the environment image information, the weight variation and the action track of the live pig.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110350581.1A CN113068657B (en) | 2021-03-31 | 2021-03-31 | Intelligent efficient pig raising method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110350581.1A CN113068657B (en) | 2021-03-31 | 2021-03-31 | Intelligent efficient pig raising method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113068657A true CN113068657A (en) | 2021-07-06 |
| CN113068657B CN113068657B (en) | 2022-04-08 |
Family
ID=76614482
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110350581.1A Active CN113068657B (en) | 2021-03-31 | 2021-03-31 | Intelligent efficient pig raising method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113068657B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114467798A (en) * | 2022-02-10 | 2022-05-13 | 重庆市畜牧技术推广总站 | Intelligent pig raising growth health dynamic monitoring management method based on big data |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101144705A (en) * | 2007-07-25 | 2008-03-19 | 中国农业大学 | Method for monitoring pig growth using binocular vision technology |
| GB201400288D0 (en) * | 2014-01-08 | 2014-02-26 | Greengage Lighting Ltd | A method of livestock rearing and a livestock shed |
| CN109141248A (en) * | 2018-07-26 | 2019-01-04 | 深源恒际科技有限公司 | Pig weight measuring method and system based on image |
| CN110612921A (en) * | 2019-09-25 | 2019-12-27 | 农芯科技(广州)有限责任公司 | Monitoring system and method for positioning gilts |
| CN111011247A (en) * | 2019-12-13 | 2020-04-17 | 巨星农牧股份有限公司 | Sustainable and intelligent pig raising method |
| CN111294565A (en) * | 2020-03-10 | 2020-06-16 | 福建农业职业技术学院 | Intelligent pig raising monitoring method and management terminal |
| CN111670834A (en) * | 2020-07-11 | 2020-09-18 | 霍祥明 | Big data-based intelligent detection and management system for pig breeding and growth health |
| CN112070621A (en) * | 2020-09-09 | 2020-12-11 | 新希望六和股份有限公司 | Pig feeding management method, pig feeding management system, computer equipment and storage medium |
-
2021
- 2021-03-31 CN CN202110350581.1A patent/CN113068657B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101144705A (en) * | 2007-07-25 | 2008-03-19 | 中国农业大学 | Method for monitoring pig growth using binocular vision technology |
| GB201400288D0 (en) * | 2014-01-08 | 2014-02-26 | Greengage Lighting Ltd | A method of livestock rearing and a livestock shed |
| CN109141248A (en) * | 2018-07-26 | 2019-01-04 | 深源恒际科技有限公司 | Pig weight measuring method and system based on image |
| CN110612921A (en) * | 2019-09-25 | 2019-12-27 | 农芯科技(广州)有限责任公司 | Monitoring system and method for positioning gilts |
| CN111011247A (en) * | 2019-12-13 | 2020-04-17 | 巨星农牧股份有限公司 | Sustainable and intelligent pig raising method |
| CN111294565A (en) * | 2020-03-10 | 2020-06-16 | 福建农业职业技术学院 | Intelligent pig raising monitoring method and management terminal |
| CN111670834A (en) * | 2020-07-11 | 2020-09-18 | 霍祥明 | Big data-based intelligent detection and management system for pig breeding and growth health |
| CN112070621A (en) * | 2020-09-09 | 2020-12-11 | 新希望六和股份有限公司 | Pig feeding management method, pig feeding management system, computer equipment and storage medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114467798A (en) * | 2022-02-10 | 2022-05-13 | 重庆市畜牧技术推广总站 | Intelligent pig raising growth health dynamic monitoring management method based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113068657B (en) | 2022-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Feeding behavior recognition for group-housed pigs with the Faster R-CNN | |
| Fernandes et al. | A novel automated system to acquire biometric and morphological measurements and predict body weight of pigs via 3D computer vision | |
| Chen et al. | Image motion feature extraction for recognition of aggressive behaviors among group-housed pigs | |
| Baek et al. | Using machine vision to analyze and classify Caenorhabditis elegans behavioral phenotypes quantitatively | |
| Subedi et al. | Tracking pecking behaviors and damages of cage-free laying hens with machine vision technologies | |
| CN107330403B (en) | Yak counting method based on video data | |
| CN110532899B (en) | Sow antenatal behavior classification method and system based on thermal imaging | |
| CN115830490A (en) | Multi-target tracking and behavior statistical method for herd health pigs | |
| Wang et al. | A PCA-based frame selection method for applying CNN and LSTM to classify postural behaviour in sows | |
| CN113068657B (en) | Intelligent efficient pig raising method and system | |
| Li et al. | Image processing for analyzing broiler feeding and drinking behaviors | |
| CN114092378A (en) | Animal health detection method, device, equipment and storage medium | |
| CN111160422B (en) | Analysis method for detecting attack behaviors of group-raised pigs by adopting convolutional neural network and long-term and short-term memory | |
| CN113516139A (en) | Data processing method, device, equipment and storage medium | |
| CN112528823A (en) | Striped shark movement behavior analysis method and system based on key frame detection and semantic component segmentation | |
| CN115359050B (en) | Method and device for detecting abnormal feed intake of livestock | |
| Hitelman et al. | The effect of age on young sheep biometric identification | |
| Bastiaansen et al. | Continuous real-time cow identification by reading ear tags from live-stream video | |
| Nasiri et al. | An automated video action recognition-based system for drinking time estimation of individual broilers | |
| CN114500960B (en) | Inspection and alarm management system and method for livestock and poultry farm | |
| CN115272943B (en) | Livestock and poultry feeding abnormity identification method based on data processing | |
| CN115661717A (en) | Livestock crawling behavior marking method and device, electronic equipment and storage medium | |
| CN116798113A (en) | Method for predicting potential morbidity of livestock | |
| CN115641466B (en) | Method for discriminating sick cattle based on video data | |
| Hocke | Detecting honey bee trophallaxis in image data with convolutional neural networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |