阅读说明：本技术 一种基于人工智能对反刍动物发情行为识别的方法 (Method for recognizing oestrus behavior of ruminant based on artificial intelligence ) 是由 彭东乔 刘炜 于 2020-06-10 设计创作，主要内容包括：本发明公开了一种基于人工智能对反刍动物发情行为识别的方法,涉及反刍动物发情监测技术领域,其包括如下步骤：A.从俯视角度拍摄反刍动物的图片,并对图片中的反刍动物进行标记；B.利用少样本学习方法对图片进行处理,建立评估预测模型；C.利用摄像设备采集反刍动物的视频资料,并对视频资料中的反刍动物进行标记,然后利用评估预测模型对视频资料中的反刍动物行为进行分析；D.根据分析结果确定反刍动物是否发情。本发明提供的基于人工智能对反刍动物发情行为识别的方法能同时对大量反刍动物进行监测,且相对于现有方法而言能大幅度提高识别的准确率,降低使用成本。(The invention discloses a method for identifying the oestrus behavior of a ruminant based on artificial intelligence, which relates to the technical field of monitoring the oestrus of the ruminant and comprises the following steps: A. shooting a picture of a ruminant from a top view angle, and marking the ruminant in the picture; B. processing the picture by using a few-sample learning method, and establishing an evaluation prediction model; C. the method comprises the steps that video data of ruminants are collected through camera equipment, the ruminants in the video data are marked, and then the behavior of the ruminants in the video data is analyzed through an evaluation prediction model; D. and determining whether the ruminant is in heat according to the analysis result. The method for recognizing the oestrus behavior of the ruminant based on artificial intelligence can monitor a large number of ruminants simultaneously, and compared with the existing method, the method can greatly improve the recognition accuracy and reduce the use cost.)

1. A method for recognizing oestrus of ruminants based on artificial intelligence is characterized by comprising the following steps:

A. shooting a picture of a ruminant from a top view angle, and marking the ruminant in the picture;

B. processing the picture by using a few-sample learning method, and establishing an evaluation prediction model;

C. the method comprises the steps that video data of ruminants are collected through camera equipment, the ruminants in the video data are marked, and then the behavior of the ruminants in the video data is analyzed through an evaluation prediction model;

D. and determining whether the ruminant is in heat according to the analysis result.

2. The artificial intelligence based ruminant oestrus behavior recognition method as claimed in claim 1, wherein the ruminant is marked by the following method: the method comprises the steps of marking three positions of the head, the neck and the hip of the ruminant, and simultaneously marking the identity of the ruminant.

3. The artificial intelligence based ruminant oestrus behavior recognition method as claimed in claim 1, wherein the evaluation prediction model outputs an analysis result of a first type or a second type, or outputs the first type and the second type simultaneously, wherein the first type is: the number of climbing spans of a ruminant in one day and the duration of each climbing span; the second method is as follows: distance moved by the ruminant per unit time.

4. The method for identifying the oestrus behavior of a ruminant according to claim 3, wherein whether the ruminant climbs or not is judged according to the relative position of the head, the neck and the hip, and when the connecting line of the neck and the hip of the ruminant is overlapped with the connecting line of the head, the neck and the hip of another ruminant, the ruminant can be judged to climb.

5. The artificial intelligence based ruminant oestrus behavior recognition method according to claim 4, wherein the ruminant's movement distance is calculated by:

selecting any one point of head, neck and hip of ruminant as reference point, and moving distance X when ruminant moves linearly₁Is the straight line between the positions of the reference point before and after movementA line distance;

when the ruminant is turned to move, a connecting line of the head, the neck and the hip of the ruminant before moving is marked as a reference line, a connecting line of the head, the neck and the hip of the ruminant after moving is marked as a judgment line, a perpendicular line intersected with the reference line is made from a reference point on the judgment line, and the length of the perpendicular line is marked as X₁' the length between the foot and the reference point on the reference line is denoted X₂' then X₁＝X₁＇+X₂＇。

6. The artificial intelligence based ruminant oestrus behavior recognition method as claimed in claim 5, wherein the two methods for determining whether the ruminant is oestrus in step D are as follows:

the method comprises the following steps: determining a time length T as a unit time, and monitoring the activity of the ruminant in the unit time to obtain the activity T of the ruminant in the unit time₀(ii) a Collecting data m days before monitoring as data storage, calculating average activity mean in unit time of m days before monitoring, and activity standard deviation STD of m days before monitoring₀If T is₀>mean+nSTD₀Determining that the ruminant is in an oestrus state;

wherein n is an adjusting coefficient, n is more than or equal to 1, and n is a natural number;

the second method comprises the following steps: if b is₁B₁+b₂B₂+b₃B₃+b₄B₄+b₅B₅+b₆B₆>If the animal is 100, the ruminant is determined to be in estrus;

if b is₁B₁+b₂B₂+b₃B₃+b₄B₄+b₅B₅+b₆B₆>If the animal is 50, determining that the ruminant is suspected to be in estrus;

wherein, B₁＝5，B₁Intermittent movement of the ruminant occurs in a unit time; b is₂＝10，B₂Is that the ruminant is climbed but stands for less than 3 s; b is₃＝10，B₃Is reversedThe ruminant is in contact with the animal but does not climb across; b is₄＝15，B₄Placing a chin rest on the body of the other animal for the ruminant; b is₅＝35，B₅Attempting to climb across other animals for ruminants; b is₆＝100，B₆Receiving climb and standing immobility time for ruminant>＝3s；b₁Is a unit time B₁Accumulating the occurrence times of the corresponding behaviors; b₂Is a unit time B₂Accumulating the occurrence times of the corresponding behaviors; b₃Is a unit time B₃Accumulating the occurrence times of the corresponding behaviors; b₄Is a unit time B₄Accumulating the occurrence times of the corresponding behaviors; b₅Is a unit time B₅Accumulating the occurrence times of the corresponding behaviors; b₆Is a unit time B₆The number of occurrences is accumulated for the corresponding behavior.

7. The method for identifying the oestrus of the ruminants based on artificial intelligence as claimed in claim 6, wherein the step D can be used for identifying the oestrus of the ruminants by adopting a method I or a method II independently, or can be used for mutual authentication and identification by adopting the method I or the method II simultaneously.

8. The method for identifying the oestrus behavior of a ruminant according to claim 1, wherein in the step C, the ruminant is marked by using a Supervisory platform when video data of the ruminant is collected.

9. The artificial intelligence based ruminant oestrus behavior recognition method as claimed in claim 1, wherein the ruminant comprises breeding cows, goats and sheep.

Technical Field

The invention relates to the technical field of ruminant oestrus monitoring, in particular to a method for recognizing ruminant oestrus behaviors based on artificial intelligence.

Background

In the operation of breeding farms for ruminants (calves, lambs) and dairy farms (cows, goats), breeding management is an important link. Taking cattle as an example, the ideal breeding cycle of a common cow is 1 fetus in 1 year, including a pregnancy period of 280 days and a nonpregnant period of 30-60 days. After the non-pregnancy period is finished, the prepared cow is subjected to artificial insemination during estrus, and the next round of breeding is carried out. The average estrus cycle of the cow is 21 days, and if the cow misses an estrus, the cow waits for the next estrus cycle for artificial insemination. Therefore, it is important to accurately detect the oestrus of cows in time in order to maintain a normal production level continuously.

For a breeding farm, if the oestrus of a cow is missed once, artificial insemination on a prepared cow cannot be performed in time, and only the arrival of the next oestrus cycle can be waited; for a cow farm, if oestrus cannot be detected accurately in time after a cow is born and a rest (generally 30 to 60 days), the number of days of nonpregnant increases. The feed cost is increased by adding the nonpregnant days for preparing the cows every day. The larger the number of cows raised, the greater the economic losses that result therefrom.

The expression of cow oestrus can be summarized into two categories: primary and secondary markers. The main sign is standing estrus, i.e. an estrus cow receives climbing (mouting) of other cows and stands still. The secondary markers include behavioral characteristics that the heat cow only climbs other cows, vulva is red and swollen, mucus secretion amount is increased, sitting and standing are uneasy, activity amount is greatly increased and the like. Therefore, the timely and accurate detection of the oestrus of the cows is the key for successful mating.

In recent years, with the advancement of science and technology, a plurality of tools and devices for detecting the estrus of cows are successively introduced, such as an estrus detector for detecting the estrus of cows by pressure, an estrus detecting system for recognizing the estrus by sound and body temperature, and the like. The systems provide technical support for oestrus detection of cows, but the problem defects are obvious, one is that real-time detection cannot be achieved by a plurality of detection systems, and the operation cost of a farm rises linearly along with increase of the total number of raised cows, so that the systems are difficult to popularize and use in the farm in a large area due to the defects.

However, the most used methods up to now are still manual observation and pedometer methods of the cow as a whole. The accuracy of the manual observation method is between 50% and 70%, and the method is usually set to a fixed time for observation in one day due to the labor cost, so that the cow oestrus which is carried out outside the observation time is difficult to detect, and the oestrus detection rate of the whole cow is large. And with the increase of the total number of the fed cows, the difficulty of comprehensively carrying out a manual observation method is further increased, and the detection rate is further reduced. The pedometer method is to detect the activity of cow by combining with a pedometer auxiliary system and calculate the estrus of cow according to the change of the activity. Previous studies show that the accuracy of the monitoring method based on the pedometer auxiliary system is between 51% and 87%, but each cow needs to be provided with independent equipment, and the larger the total number of raised cows is, the heavier the burden on the operation of a farm is.

Therefore, the cow oestrus identification method can keep high cow oestrus identification accuracy, inform the oestrus state of each cow to farm workers in time, and be conveniently used in the farm at a low price, and the identification detection scheme is particularly important in sustainable development of the farm.

Disclosure of Invention

In order to solve the problems, the invention provides a method for identifying the oestrus behavior of the ruminant based on artificial intelligence, which monitors and analyzes the primary mark and the secondary mark of the oestrus of the cow simultaneously, and greatly improves the comprehensiveness and the accuracy of oestrus monitoring.

The invention specifically adopts the following technical scheme for realizing the purpose:

a method for recognizing oestrus behavior of ruminants based on artificial intelligence comprises the following steps:

A. shooting a picture of a ruminant from a top view angle, and marking the ruminant in the picture;

B. processing the picture by using a few-sample learning method, and establishing an evaluation prediction model;

C. the video data of the ruminants are collected by using camera equipment, the ruminants in the video data are marked, and then the behavior of the ruminants in the video data is analyzed by using an evaluation prediction model;

D. and determining whether the ruminant is in heat according to the analysis result.

Further, the labeling method of the ruminant is as follows: the method comprises the steps of marking three positions of the head, the neck and the hip of the ruminant, and simultaneously marking the identity of the ruminant.

Further, the analysis result output by the evaluation prediction model is a first type or a second type, or the first type and the second type are output simultaneously, wherein the first type is: the number of climbing spans of a ruminant in one day and the duration of each climbing span; the second method is as follows: distance moved by the ruminant per unit time.

Furthermore, whether the ruminant climbs or not is judged according to the relative positions of the head, the neck and the hip of the ruminant, and when the connecting line of the neck and the hip of the ruminant is overlapped with the connecting line of the head, the neck and the hip of another ruminant, the ruminant can be judged to climb.

Further, the moving distance of the ruminant is calculated by:

selecting any one point of head, neck and hip of ruminant as reference point, and moving distance X when ruminant moves linearly₁The linear distance between the positions of the reference point before and after the movement is taken as the reference point;

when the ruminant is turned to move, a connecting line of the head, the neck and the hip of the ruminant before moving is marked as a reference line, a connecting line of the head, the neck and the hip of the ruminant after moving is marked as a judgment line, a perpendicular line intersected with the reference line is made from a reference point on the judgment line, and the length of the perpendicular line is marked as X₁' the length between the foot and the reference point on the reference line is denoted X₂' then X₁＝X₁＇+X₂＇。

Further, the method for judging whether the ruminant is in heat in the step D includes the following two methods:

the method comprises the following steps: determining a time length t as a unit time, monitoring the activity of the ruminant in the unit timeMeasuring to obtain the activity T of the ruminant in the unit time₀(ii) a Collecting data m days before monitoring as data storage, calculating average activity mean in unit time of m days before monitoring, and activity standard deviation STD of m days before monitoring₀If T is₀>mean+nSTD₀Determining that the ruminant is in an oestrus state;

wherein n is an adjusting coefficient, n is more than or equal to 1, and n is a natural number;

the second method comprises the following steps: if b is₁B₁+b₂B₂+b₃B₃+b₄B₄+b₅B₅+b₆B₆>If the animal is 100, the ruminant is determined to be in estrus;

if b is₁B₁+b₂B₂+b₃B₃+b₄B₄+b₅B₅+b₆B₆>If the animal is 50, determining that the ruminant is suspected to be in estrus;

wherein, B₁＝5，B₁Intermittent movement of the ruminant occurs in a unit time; b is₂＝10，B₂Is that the ruminant is climbed but stands for less than 3 s; b is₃＝10，B₃The ruminant is in contact with the animal but does not climb; b is₄＝15， B₄Placing a chin rest on the body of the other animal for the ruminant; b is₅＝35，B₅Attempting to climb across other animals for ruminants; b is₆＝100，B₆Receiving climb and standing immobility time for ruminant>＝3s；b₁Is a unit time B₁Accumulating the occurrence times of the corresponding behaviors; b₂Is a unit time B₂Accumulating the occurrence times of the corresponding behaviors; b₃Is a unit time B₃Accumulating the occurrence times of the corresponding behaviors; b₄Is a unit time B₄Accumulating the occurrence times of the corresponding behaviors; b₅Is a unit time B₅Accumulating the occurrence times of the corresponding behaviors; b₆Is a unit time B₆The number of occurrences is accumulated for the corresponding behavior. .

Furthermore, the step D may separately adopt the first method or the second method to identify the oestrus behavior of the ruminant, or may simultaneously use the first method and the second method to perform mutual authentication identification.

Further, in the step C, when the video data of the ruminant is collected, the ruminant may be marked by using a Supervisely platform.

Still further, the ruminants include breeding cows, goats, and sheep.

The invention has the following beneficial effects:

1. the method takes the positions of the head, the neck and the hip of the ruminant as key points for judging whether the crawling behavior exists or not, simplifies a plurality of key points compared with other judging methods, greatly reduces the operation amount when deep learning prediction is used, and can identify the estrus behavior of the cattle in real time only under the condition of marginal deployment without using cloud service; the ruminant breeding is generally carried out in remote mountain areas or rural areas, the network conditions of the areas are poor, and the method has good adaptability and extremely high practical significance for ruminant breeding;

2. in the invention, besides using few samples for deep learning when extracting key points, simple calculation is only used for judging the crawling and crossing behaviors of the ruminant and calculating the moving distance of the ruminant, thereby greatly improving the edge prediction speed of the whole system and reducing the memory requirement of the system;

3. according to the invention, a large number of experiments are carried out to obtain the correlation between the climbing times and duration, the moving distance of the ruminant and whether the ruminant is oestrous, so that the oestrus condition of the ruminant can be judged more accurately, compared with the method, most of the existing judging methods only pay attention to whether the ruminant has the climbing behavior or not, or only pay attention to the number of steps taken by the ruminant, the method combines the climbing behavior and the walking behavior, and multiple tests show that the accuracy of the identification method is up to 85% -95%, and the comprehensiveness and the accuracy of oestrus monitoring of the ruminant are greatly improved;

4. the implementation cost of the invention is low, the monitoring range is large, the ruminant oestrus monitoring cost can be effectively reduced, and the farm income is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

FIG. 1 is a schematic view of a cow climbing state;

FIG. 2 is a schematic diagram of a cow cheek rest state;

FIG. 3 is a schematic diagram of a linear movement distance calculation;

FIG. 4 is a schematic view of a steering movement distance calculation;

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "upper", "lower", "inner", "outer", and the like refer to the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship which the product of the present invention is conventionally placed in use, and are used for convenience of description and simplification of description, but do not refer to or imply that the device or element referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and the like are used merely to distinguish one description from another, and are not to be construed as indicating or implying relative importance.

In the description of the present invention, it should also be noted that, unless otherwise explicitly specified or limited, the terms "disposed" and "in communication" are to be interpreted broadly, e.g., as either fixed or removable communication, or integrally connected; either mechanically or electrically; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.