Detecting estrus in animals for insemination

Abstract

A method for detecting estrus of sows where insemination is performed at the time a determination is made. In a first method, an image of the hind profile of each sow is analyzed to determine whether the sow is in estrus from characteristics related to the shape, size, and / or color of the vulva. In this analysis, a program may be used to pre-analyze multiple images showing the vulva areas of multiple sows known to be in estrus and multiple sows known not to be in estrus to generate data of the previous images, whereby a program is written to generate a signal using the data to make the determination. In a second method, a boar is confined and placed in close contact with the sows, and images and / or sounds of each sow are analyzed to determine whether the sow is in estrus.

Description

[Technical field]

[0001] The present invention relates to a method for detecting estrus in animals for use in insemination. [Background technology]

[0002] For efficient animal husbandry, much attention has been paid to finding effective methods for determining when the animals concerned are in estrus and thereby increasing the chances of successful insemination.

[0003] It is known that changes in movement and demeanor, including changes in the ratio of periods standing to periods lying down, can be used to detect estrus. Additionally, temperature of the body or particularly the vaginal area can be used to make the determination.

[0004] However, further improvements in the techniques for making such determinations are needed.

[0005] The present invention is particularly applicable to sows, where personnel typically have less contact with the animals and therefore require a more automated method that can provide a determination of estrus, however, the method can also be used with other animals. Summary of the Invention

[0006] The object of the present invention is to provide a method for raising livestock, the method comprising the steps of: - defining a containment area in which a group of sows is restricted, where the sows are potentially in estrus; Providing a confined or restrained boar; The containment area and the restricted space are configured such that the sow in the containment area can come into close contact with the boar by movement of the boar or the sow; providing at least one collection device configured to collect images and / or sounds of each sow when in said intimate contact; and using a control system to analyze the images and / or sounds to determine if the sow is in estrus for insemination.

[0007] Thus, the methods herein may use images from a camera when the sow is in close contact with the boar, or may use sounds produced by the sow, or may combine analysis of both data sets.

[0008] In a preferred method, images are taken from the front of the sow to show the changes in body and head movement and appearance that are indicative of estrus. In particular, frontal elevation images of the sow can be taken from the front when the sow is close to the front of the boar where the changes are expected to be most evident.

[0009] In a preferred method, the images are analyzed for sow movements characteristic of estrus, as determined by previous analyses of sows known to be in estrus, and compared to the currently acquired image. To obtain more accurate results, errors can be eliminated and multiple comparisons can be made using statistical analysis.

[0010] Specifically, the images may be analyzed for features known to be indicative of estrus, including: The speed of movement characteristic of estrus, Characteristic body movements during estrus Head movements characteristic of estrus The characteristic ear movements during estrus Angular velocity of ears characteristic of estrus, The dorsal region movement characteristic of estrus, This includes the changes in body posture characteristic of estrus.

[0011] Specifically, the sounds may be analyzed for features known to be indicative of estrus, including: Certain types of sounds made by sows that are characteristic of estrus, such as squealing and grunting, The spectrum of frequencies characteristic of the estrus period, It includes a series of sounds characteristic of the rutting season.

[0012] Preferably the collection device comprises a camera attached to an installation which restrains or confines the boars and which is aimed at the sows in the installation.

[0013] In one embodiment, the restraining or confining equipment includes a mobile structure where boars are transported to sows while the sows are in their individual containment pens. This is typically a boar cart, where boars are moved on a support floor along the path to sows who are confined in cages and confined to crates along the path. Boar carts provide the best way to confine and control large boars as they move from sow to sow. However, other means for transporting boars can be used, such as simple tethering or guided robotic devices, where boars are restrained while walking from sow to sow.

[0014] In an alternative configuration, the equipment for restricting or confining boars is fixed in place as fixed cages within an open confinement area where sows can move freely from place to place, up to and including boars which are separately confined in their own cages within the confinement area.

[0015] In both cases the camera is preferably positioned on a cart or cage facing the sow to be photographed so that a series of images of the sow are obtained as the sow approaches or is approached, and those images are used in the analysis of the sow's movements as defined above.

[0016] Preferably, the microphone is of a type that is further positioned in the boar's cage and oriented to receive primarily sow sounds in the boar's confined space.

[0017] Although the above description uses the term "boar", it is understood that this can be an actual live animal. However, the boar can also be a manufactured structure that provides one or more stimuli to the sow that simulate a live animal. In the extreme case, the "boar" can simply be the use of a pheromone spray to adequately simulate a boar. Live animals have been found to be more effective in stimulating sows to detect estrus and to accept insemination. However, improvements in the simulation stimuli can avoid the need to transport or cage actual live, very large animals.

[0018] In an important optional feature that may be used with any of the above descriptions or with any of the features defined below, the control system is operative to distinguish between determinations of estrus requiring simple and complex analysis, where simple analysis provides an output signal indicative of a determination of the sow's estrus, and complex analysis provides an output signal requiring additional intervention before a determination of estrus is made. That is, the system may generate and indicate a level of confidence in the result, whereby the system may proceed directly to a determination or indicate a low confidence requiring further data to be provided. For example, the further data may come from the methods described below or from other intervention by the operator. As another example, the operator may be employed to apply pressure to the sow's dorsal area during collection of sounds and / or images.

[0019] The result of the determination is that insemination is performed upon receipt of a signal indicating that a final determination has been made, such that the probability of success of the insemination procedure is increased.

[0020] In another method, which can be used as a separate system or as part of the above analysis, a method for raising livestock is provided herein, the method comprising: defining a containment area within which a group of potentially estrus sows is confined; periodically collecting images of the hind profile of each sow; analysing each image to determine an area on the image representing the vulva of the sow; analyzing the images of the area using a control system to determine whether the sow is in estrus based on characteristics relating to vulva shape, size, and / or color; providing an output signal in response to determining that the sow is in estrus; Insemination may be performed upon receipt of a signal indicating that a determination has been made.

[0021] Determining if a sow is in heat can be used for timed insemination or simply to know the animal has gone into heat. In some cases, farmers may want to skip a estrous cycle for some sows to optimize reproductive outcomes.

[0022] In a preferred method, the analysis is performed by the control system using a program where a program is created that pre-analyzes a number of images showing the vulva areas of a number of sows known to be in estrus and a number of sows known not to be in estrus to generate data for the previous images, whereby the data is used to generate a signal to make the determination. Thus, the program that operates the control system is a neural network that has pre-trained features to determine the presence of estrus by analyzing a number of images of the vulva areas of animals known to be in estrus and animals known not to be in estrus. This pre-trained programming can be stored in the control unit, which may be on-premise or on the cloud, so that it is applied to new images to be analysed.

[0023] While the above machine learning is performed by the system itself, the applicant has found that a broader analysis of the image can allow the determination to be made, for example, based on a number of visual features that can be extracted from the image. These features can include one or more of the following: -a- Vulvar tissue or wrinkles; this relates to the degree of swelling of the area as the absence of wrinkles occurs due to an increase in volume caused by blood congestion which stretches the skin and reduces the depth and expression of the wrinkles. -b- Height or width of the vulva; again, this relates to the degree of expansion of the area, as the height and width of the area increases with swelling. -c- Vulva color; the presence of concentrated blood acts to intensify the red color that can be observed. -d- Outer prominence of the vulva. Again, engorgement of the area due to increased blood supply acts to expand the vulva outwardly towards the body, producing an increased 3D effect that can be observed even from a 2D image when photographed or selected at a suitable angle.

[0024] In some cases, especially in older animals that have produced many litters, the vulva area may be less obvious in changes, so that analysis of a single image may be inconclusive. In this case, comparison with previous images from the same animal may be used to observe changes over time, where at least one image of the area is compared with a previous image of the area of ​​the same sow, and the difference over time is used to make a determination. That is, in many sows the changes are sufficiently obvious that the presence of estrus can be detected from a single image or a series of images taken at the same time, which represent the current state at the time of the test. It will be understood that the images, and the determination from the images, are carried out repeatedly over a period of time, for example daily, in order to make a determination as soon as estrus begins. Thus, a series of images is available over that period for comparison. In other cases, comparison with older images is necessary to make a determination. However, this of course requires that the images be associated with the identity of the animal concerned, which is easily done when the animal is confined, but additional measures are required, especially when the animal is allowed to move freely.

[0025] In the first step of the analysis, preferably, multiple images of the posterior profile are obtained and analyzed to select one or more suitable images that best show the area. If the camera is mounted and positioned so that it can change its angle and position relative to the rear of the animal, many of the images of the posterior profile may be inappropriate to show the vulva area, and therefore the control system uses a program to select the images that are appropriate and use them for the determination. As part of this analysis, the system needs to select the relevant vulva area from the entire image of the posterior profile, so that part of the algorithm requires that this area is first determined and selected to be the image component for further analysis. That is, after the vulva area is selected, it is necessary to select from that subset of image components the component that best shows the area and should be used for the determination. Each image is then evaluated by the system to determine whether it shows a positive or negative result. Following the determination, a comparison of all positive and negative results will give a final determination and a confidence level, which is indicated by how many results match. This information can be used to assess whether insemination will occur, i.e. whether the vulva shows signs of estrus, or whether further analysis with additional data is required.

[0026] After a first analysis to select from the posterior contour images a number in the subset that best or preferably represent the vulva region, these multiple images of the region are selected and compared to arrive at a determination using statistical analysis.

[0027] In some embodiments, sows are confined to individual stalls and images are obtained by cameras mounted in the stalls. In this case, the cameras can be mounted in fixed positions in the stalls and can be timed and triggered to repeatedly take images of the rear profile at a frequency determined by the system, for example daily. The cameras can then take multiple images of the rear profile for the above analysis, taking into account that the animals may not be in the right position or at the right angle relative to the camera to obtain the best rear profile image to be obtained. In other cases, these images may not all be suitable either, as the images are taken by a moving camera (on a rail, on a robot, etc.).

[0028] Preferably, the rear profile images are taken while the animal is standing to ensure that the vulva region is properly presented posteriorly and to allow easier selection by locating the vulva region at a specific location within the image. Furthermore, such a standing position ensures that portions of the region are consistently presented in the image allowing for more effective analysis.

[0029] It may therefore be desirable to ensure that the animal is standing when the image is acquired, or even to induce the animal into a standing position with some physical cue in order to obtain the image.

[0030] In some embodiments, sows are not confined to individual compartments so that they are free to move around. In this case, images can be obtained by cameras mounted in suitable locations, such as feeding or watering compartments, frequented by sows. This can also be done in boar exhibition stations, where there are openings through which sows can view boars and obtain sexual stimulation. In this way, each sow can be moved randomly to the camera and imaged each time it moves to the relevant position. In one configuration, in this situation, it is necessary to apply an ID to the animal to be imaged so that the animal can be identified for insemination. Therefore, in this case, the animal can be tagged using a readable tag, such as an RFID ear tag. This ID can also be used for the historical comparison described above.

[0031] Alternatively, if ID information is not available, after grading, the sows to be graded for later insemination can be marked. Such marking systems with paint sprays or dabbers are available and can be activated to mark the relevant sows. This is particularly useful when the imaging camera is positioned in a feeding area with walls that contain one animal at a time so that it can be contained, imaged in a suitable orientation for imaging, and graded before returning the animal to the group. Alternatively, a pathway may be provided between two zones, such as a sleeping zone and a feeding zone, where the animals are temporarily held to complete an action before proceeding to the next zone.

[0032] In some cases, the system may be provided as a program that can be applied to existing imaging systems, such as mobile phones. In this case, images can be taken manually by an operator holding an imaging device, who can provide information on the animal concerned, store the data and select the images to be taken. From those images, the control system takes over and selects the area images to be used and makes a decision from the multiple images to be communicated to the operator. In other cases, the control system may be part of a stand-alone system that includes a camera, a processor for making the decision and an interface for communicating with the operator. In both cases, the control system may be configured to provide data regarding the decisions of multiple sows to the sow management system.

[0033] As described in detail above, the image-based methods may be used in conjunction with providing a confined or restricted boar, where the sow containment area and boar containment system are configured such that sows within the containment area are able to come into close contact with the boar by movement of the boar or sow. [Brief description of the drawings]

[0034] One embodiment of the present invention is described below in conjunction with the accompanying drawings. [Figure 1] FIG. 1 is a schematic plan view of an animal breeding barn showing a configuration used in a first embodiment of the method according to the present invention. [Diagram 2] FIG. 2 is a schematic plan view of an animal breeding barn showing a configuration used in a second embodiment of the method according to the present invention. [Diagram 3] 1 is a flow chart showing the steps of a first embodiment of a method according to the present invention. [Figure 4] FIG. 11 is a schematic plan view of an animal breeding barn showing a configuration used in a third embodiment of the method according to the present invention. [Diagram 5] FIG. 11 is an example of an image of the posterior ridge of an animal used in the third embodiment of the method according to the present invention. [Figure 6]FIG. 6 is an example of an area image of an animal taken from FIG. 5 and used in a third embodiment of the method according to the invention. [Figure 7] 4 is a flow chart showing the steps of a third embodiment of the method according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0035] Components used in a method for raising livestock are shown in Figure 1. A confinement area (10) includes multiple side-by-side breeding pens (11) for confined sows arranged in a row along paths (14).

[0036] The pen is formed by welded bar pen walls with a feed trough (12) along the front wall so that the animals remain confined during the breeding season when the herd is restricted to sows potentially in estrus.

[0037] The boars are confined to a movable boar cart (15) on wheels (16) and movable along a path (14) which in turn brings them to each sow, allowing close contact with the boars as the boars move. Suitable carts are shown in prior USP 6,196,975 (Labrecque), issued March 6, 2001, and USP 6,508,202 (Kleinsasser), issued January 21, 2003, both of which are assigned to the present applicant, the disclosures of which are incorporated herein by reference.

[0038] The cart carries a camera (17) and a microphone (18) at fixed positions on the cart, the camera taking images and the microphone picking up sounds of each sow when in close contact, so the camera and microphone are placed in appropriate positions on the frame of the cart and oriented to the required position when the cart is registered to the sow pen.

[0039] The images and sounds are transmitted by wire or wirelessly to a control system (19) which analyzes the data as described below to determine whether the sow is in estrus for insemination.

[0040] As shown in the flow chart of Figure 3, the method can use either images from a camera or sounds emitted by the sow, or combine analysis of both data sets, where images and sounds are acquired when the sow is in close contact with the boar. The steps of the method are detailed in Figure 3.

[0041] In a preferred method, images are taken from the front of the sow for a frontal view showing the body and head as they exhibit the movements and attitude changes that are determinative of estrus. Frontal view images of the sow can be taken from the front, particularly when the sow is in close proximity to the front of the boar, where the changes are expected to be most evident.

[0042] As shown in Figure 3, the images are analyzed for sow movements characteristic of estrus, determined by previous analysis of sows known to be in estrus, and compared to the currently acquired image or images with material from a database. To obtain more accurate results, errors can be eliminated and multiple comparisons can be made using statistical analysis.

[0043] Specifically, the images may be analyzed for features known to be indicative of estrus, including: The speed of movement characteristic of estrus, Characteristic body movements during estrus Head movements characteristic of estrus The characteristic ear movements during estrus Angular velocity of ears characteristic of estrus, The dorsal region movement characteristic of estrus, This includes the changes in body posture characteristic of estrus.

[0044] Specifically, the sounds may be analyzed for features known to be indicative of estrus, including: Certain types of sounds that sows make that are characteristic of estrus, such as squealing and grunting, The spectrum of frequencies characteristic of the estrus period, It contains a series of sounds characteristic of the rutting season.

[0045] As shown in FIG. 2, the sow is free in the containment pen (20) when she wishes to move to the feeding (21), watering (22) or sleeping areas. The containment pen for the sow is fixed in place as a fixed cage in an open containment area, where the sow is free to move from place to place, including the boar, which is individually confined in its own cage in the containment area. In this case, a camera (24) and microphone (25) are placed on the cage in a position facing the sow as the boar approaches the boar cage from in front of the boar. In this way, a series of images of the sow can be obtained as it approaches, in order to use these images in the analysis of the sow's movements, as shown in FIG. 3. In FIG. 2, short dividing walls (25A) and (25B) are provided extending forward from the boar cage on either side of the sow when standing in front of the boar, so that signals from only one animal are obtained and analyzed, separated from other animals in the same general area.

[0046] The microphone (25) is an elongated directional microphone primarily for receiving direct sound from the sow in the boar's confined space and discarding sound from other areas of the containment area, such as other sows.

[0047] As shown in Figure 3, the control system operates to distinguish between estrus determinations requiring simple analysis and complex analysis, where an output signal is provided indicating a determination of the sow's estrus, and where a complex analysis provides an output signal requiring additional intervention before an estrus determination is made. That is, the system can generate and indicate a level of confidence in the result, whereby the system can either proceed directly to a determination or indicate a need to provide more data with less confidence. For example, the more data can come in response to a request from the methods described below or from other intervention by the operator. As another example, the operator may be employed to apply pressure to the sow's dorsal area during collection of sounds and / or images.

[0048] The result of the determination is that insemination is performed upon receipt of a signal indicating that a final determination has been made, such that the probability of success of the insemination procedure is increased.

[0049] In another method, shown in Figures 4-7, a containment area (30) is provided in which a group of sows are restricted, where the sows are potentially in heat. The containment area (30) includes separate pens (31) and (32) where the sows can enter to consume feed from a trough (33) or from a water supply (34). The pens are typically configured such that sows can enter alone and other sows are excluded from entering the rear of the pen. Each pen carries a fixed camera (35) which is operated to periodically collect images of the rear profile of each sow as it enters the pen. As all sows consume food periodically, all sows are subject to the imaging system during the day. In this way, each sow may be randomly moved to the camera (35) and imaged each time it moves to a position of interest. In one configuration, in this situation, it is necessary to apply an ID to the animals being imaged so that the animals can be identified for insemination. Thus, in this case, the animals may be tagged using readable tags such as RFID ear tags that are detected by a tag reader (37) in the enclosure (31).

[0050] As shown in the flow chart of FIG. 7, in which the steps are detailed, the control system goes through a series of steps as shown in order to receive the signal from the camera, analyze the images and make a determination of estrus.

[0051] Thus, each image of the posterior profile shown in FIG. 5 is analyzed using a control system such as that shown in FIG. 7 to determine the area on the image which represents the sow's vulva, as shown in FIG.

[0052] In some cases, the flow chart may vary slightly based on the method used to analyze the images. One method may not require selection from a larger image of the vulva area before selecting an appropriate image. Selection of an appropriate image for analysis may be based on other criteria depending on the angle at which the image was taken, the orientation of the animal when imaged, etc., to ensure use of the best image for analysis.

[0053] Thus, the image of the vulva region in FIG. 5 is analyzed to determine whether the sow is in estrus for insemination from features related to the shape, size, and / or color of the vulva. As above, this can be done after suitable images have been selected to extract the relevant regions from the larger image. Alternatively, the entire image may be analyzed without selecting a region. Images suitable for analysis may be selected to remove some that are inappropriate due to the angle or orientation or position of the animal, depending on how the images are obtained.

[0054] An output signal is generated indicating that a determination has been made and insemination is performed upon receipt of the signal.

[0055] In a preferred method, the analysis is performed by the control system using a program, whereby previous data of known estrus states showing the vulva areas of a plurality of sows are compared by the program. Thus, the program operating the control system is a neural network or another machine learning algorithm that has pre-trained characteristics to determine the presence of estrus by analyzing multiple images of the vulva of animals known to be in estrus, as well as animals known not to be in estrus. This pre-trained programming is stored in the control unit and applied to new images to be analyzed. Or it could even be a classical, non-machine learning algorithm, using classical computer vision techniques.

[0056] In the first step of the analysis, preferably, multiple images of the posterior contour are obtained and analyzed to select one or more suitable images that best show the area. The control system uses a program to select the images that are suitable and use them for the determination. As part of this analysis, in some cases, the system is required to select the relevant vulva area from the entire image of the posterior contour, so that part of the algorithm is required to first determine and select this area as an image component for further analysis. Each image is then evaluated by the system to determine whether it shows a positive or negative result. Following the determination, a comparison of all positive and negative results gives a final determination and a confidence level indicated by how well the results match. This information can be made available to evaluate whether insemination will occur or whether further analysis with more data is required.

[0057] One analysis that may be used to qualify the image is based on a determination by an analysis system of the cleanliness of the vulva. In this case, a warning may be given to the farmer if the vulva is too dirty and may give an erroneous result. In this case, the image may be discarded and a signal may be provided to the farmer that action is needed to clean the area before an appropriate accurate determination can be made.

[0058] In a further step of the analysis, the size determination can be based not on the actual dimensions but instead on a comparison of the size of the vulva to a determinable area of ​​the sow's hind contour, such as the size of the ruminant or the size of the total area of ​​the animal as it appears on the hind ridge or hind contour. Thus, the analysis can determine the ratio of these two areas selected by the imaging system, and the change in the detected ratio resulting from the animal progressing through estrus.

[0059] After a first analysis to select from the posterior contour images the number in the subset that is optimal or most appropriate to represent the vulva region, these multiple images of the region are selected and compared to obtain a determination using statistical analysis.

[0060] An alternative configuration is shown in Figure 1, where a camera (50) is positioned at the rear of the enclosure (11) to take images of the confined animal, either with or without the presence of a boar in a cart. Images of the rear profile are preferably taken whilst the animal is standing, to ensure that the vulva area is properly presented in the posterior direction, allowing the vulva area to be located and more easily selected at a particular location within the image. This standing position also ensures that portions of the area are consistently presented within the images, allowing for more effective analysis.

[0061] Thus, the camera (50) and associated control system are operated to acquire or select images while the animal is standing. That is, the animal can be monitored in the crate as shown in Figure 1 to determine when the animal is standing so that an image can be taken, or an actuator (51) on the crate wall is provided to move the animal by physical stimulation into a standing position for image acquisition. Alternatively, images can be taken periodically and simply deleted if the system determines that the animal is not standing.

[0062] If ID information is not available on the animals and the animals are free to roam as shown in Figure 4, after grading the sows to be graded can be marked with a marker (36) within the pen (32) for subsequent insemination. Such marking systems with paint sprays or dabbers are known and commonly available and can be operated to mark the sows concerned to indicate that they have been graded as being in estrus.

[0063] It should be noted that the timing of insemination may be offset in time from the analysis that determines the state of estrus. The actual timing of estrus may in some cases be delayed from the analysis of the vulva appearance and the detected visible changes. In fact, the vulva usually changes appearance at a time point before the standing heat reflex (conventionally used to determine estrus in sows) and sometimes continues for several days after the standing heat reflex (estrus). Thus, the appearance of the vulva is a good indicator of the presence of estrus, but not necessarily a good way to find the actual moment of insemination. However, it may be used to confirm that estrus is about to occur or is about to start. This may be followed by another system, such as the image / sound method described above, to better determine the time required for insemination. Alternatively, the vulva image system may be used to signal estrus events where it is preferable not to inseminate. Detecting these estrus events is very important in young or new sows, as farmers may wish to inseminate them in the second estrus, thus skipping the first estrus to ensure better maturation of the animals. It is therefore desirable to detect both the first and second estrus and not miss them, as otherwise the animals will be too old at their first insemination, reducing their lifetime fertility.

[0064] Because various modifications can be made in my invention as hereinabove described, and because many obviously diverse embodiments can be made within the spirit and scope of the appended claims without departing from the same, it is intended that all matter contained in the appended specification be interpreted merely as illustrative and not in a limiting sense.

Claims

A method for providing an output signal indicating that a female pig is in estrus, comprising: defining a confinement area in which a group of female pigs is restricted, wherein the female pigs in the group are potentially reaching their first estrus; periodically obtaining an image of the vulva of one female pig in the group of female pigs; analyzing the obtained vulva image using a control system to make a determination as to whether the female pig is in estrus from the image; providing a database of a plurality of images showing the vulvas of a plurality of female pigs known to be in estrus and a plurality of images showing the vulvas of a plurality of female pigs known not to be in estrus; wherein the determination is based only on a program that uses the obtained image in comparison with the database; and providing an output signal indicating that the female pig is in its first estrus in response to the determination. **Claim 2** The method according to claim 1, wherein the analysis is performed by the control system using a program, wherein a large number of images showing the vulvas of a plurality of female pigs known to be in estrus and a plurality of female pigs known not to be in estrus are pre-analyzed to generate previous image data, and thereby a program for generating a signal using the data is created for determination. **Claim 3** The method according to claim 1 or 2, wherein the determination is based on features including the tissue or wrinkles of the vulva. **Claim 4** The method according to claim 1 or 2, wherein the determination is based on features including the height or width of the vulva. **Claim 5** The method according to claim 1 or 2, wherein the determination is based on features including a comparison of the size of the vulva with a determinable area around the vulva of the female pig. **Claim 6** The method according to claim 1 or 2, wherein the determination is based on features including the color of the vulva. **Claim 7** The method according to claim 1 or 2, wherein the determination is based on features including the outer ridges of the vulva. **Claim 8** The method according to claim 1 or 2, comprising the step of analyzing the image to select an area on the image showing the vulva of the female pig for analysis. **Claim 9** The method according to claim 1 or 2, wherein a plurality of images of the vulva are analyzed to select one or more appropriate images that best show the vulva. **Claim 10**: The method according to claim 1 or 2, wherein the female pigs are confined in individual compartments, and the images are obtained by cameras attached to the compartments. **Claim 11**: The method according to claim 1 or 2, wherein the female pigs are not confined in individual compartments, and the images are obtained by cameras attached to positions frequently visited by the female pigs. **Claim 12**: The method according to claim 1 or 2, wherein a system is provided that provides identification of the female pigs from which the images were obtained. **Claim 13**: The method according to claim 1 or 2, wherein after estrus determination, the female pigs subject to the determination are marked for subsequent insemination. **Claim 14**: The method according to claim 1 or 2, wherein the control system is part of a stand-alone system including a camera, a processor for making determinations, and an interface for communicating with an operator. **Claim 15**: The method according to claim 1 or 2, wherein the control system is configured to provide data regarding the determination of multiple female pigs to a female pig management system. **Claim 16**: The method according to claim 1 or 2, wherein the images are manually taken by an operator holding an imaging device. **Claim 17**: A method according to claim 1 or 2, comprising the step of providing male pigs confined in a restricted space, wherein the confinement area and the restricted space are configured such that the female pigs in the confinement area can come into close contact with the male pigs by the movement of the male pigs or the female pigs. **Claim 18**: The method according to claim 17, comprising the step of providing at least one collection device configured to collect images and / or sounds of each female pig when in a state of close contact, and the step of analyzing the images and / or the sounds using the control system to determine whether the female pig is in estrus. **Claim 19**: The method according to claim 1 or 2, wherein the control system operates to distinguish between an estrus determination requiring simple analysis and an estrus determination requiring complex analysis, and in the simple analysis, an output signal indicating the estrus determination of the female pig is provided, and in the complex analysis, an output signal requiring additional intervention before the estrus determination is provided. The method according to claim 1 or 2, wherein the fertilization is carried out upon receipt of the output signal indicating that the determination has been made.

Images