Pig farm intelligent feeding system based on posture recognition and control method thereof

Abstract

The application relates to a pig farm intelligent feeding system based on posture recognition and a control method thereof, the system comprising a feeder, further comprising an image acquisition module, which is used for acquiring pig video images and acquiring trough video images; further comprising a pig feeding willingness judgment module, which identifies pig skeleton joints in the pig video images by using a posture recognition algorithm, judges whether a mouth joint of the pig is located within a preset trough peripheral distance and exceeds a preset time, and if yes, judges that the pig has feeding willingness, and if not, judges that the pig does not have feeding willingness; further comprising a control module, which controls feeding of the feeder according to the feeding willingness of the pig. The pig farm intelligent feeding system based on posture recognition and the control method thereof can automatically feed according to the feeding willingness of the pig, so that feeding demands of different individuals are met, the health degree of the pig is improved, and the breeding efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pig farm feeding, and in particular to a pig farm intelligent feeding system based on posture recognition and a control method thereof. BACKGROUND

[0002] The existing automatic pig feeding is generally timed automatic feeding (for example, 8:00 and 10:00 every morning), but the feeding needs of different individuals are actually different, especially for sows, this method is difficult to meet the needs, and if too much is fed, it will cause the problem of feed spoilage, affect the health of pigs, and also affect the efficiency of the pig farm. Another way is to set a pressure sensor under the feeder, when the pig arches to feed the feeder, the pressure sensor is automatically fed, but this feeding method is easy to damage the pressure sensor. The existing two feeding methods are difficult to meet the feeding needs of the pig farm. SUMMARY

[0003] In view of the technical problems in the above background art, the present application provides a pig farm intelligent feeding system based on posture recognition and a control method thereof.

[0004] In a first aspect, the present application provides a pig farm intelligent feeding system based on posture recognition, comprising a feeder, the feeder comprising a storage barrel, a feeding channel and a trough, the feeding channel connecting the storage barrel and the trough, and the trough being used to receive the food falling from the storage barrel for the pigs to eat;

[0005] An image acquisition module is configured to acquire a video image of the pig and a video image of the trough;

[0006] A pig feeding intention judgment module is configured to identify the pig skeleton joint in the pig video image by using a posture recognition algorithm, and to determine whether the mouth joint of the pig is located within a preset distance from the periphery of the trough and exceeds a preset time. If yes, it is determined that the pig has a feeding intention, otherwise, it does not have a feeding intention;

[0007] A control module is in signal connection with the pig feeding intention judgment module, and a control valve is arranged on the feeding channel, and the control valve is in signal connection with the control module.

[0008] By adopting the above technical scheme, the system can automatically feed according to the feeding intention of the pig, thereby meeting the feeding needs of different individuals, improving the health of the pig, and improving the breeding efficiency.

[0009] Preferably, the pig feeding intention judgment module is further configured to determine whether the mouth joint of the pig is blocked by the trough and exceeds a preset time. If yes, it is determined that the pig has a feeding intention, otherwise, it does not have a feeding intention.

[0010] By adopting the technical scheme, the willingness of the pig to eat can be more accurately judged.

[0011] Preferably, the image acquisition module acquires all frames in the video image by using an OpenPose posture recognition algorithm, outputs pig skeleton joints of all frames, and predicts positions of local joints that are blocked by using the OpenPose posture recognition algorithm.

[0012] Preferably, the intelligent feeding system for pig farms further comprises a control feeding amount module, which is configured to set a feeding amount according to a last feeding time of the pig.

[0013] By adopting the technical scheme, the feeding amount of the pig can be optimized, thereby facilitating resource allocation.

[0014] Preferably, the setting of the feeding amount according to the last feeding time of the pig specifically comprises: if the last feeding time of the pig is lower than a preset time, a small amount of feeding is performed, otherwise, standard feeding or overfeeding is performed.

[0015] Preferably, the intelligent feeding system for pig farms further comprises an eating judgment module, which is configured to judge whether a mouth joint of the pig coincides with or is blocked by a trough, and whether the coincidence or the blockage exceeds a preset time, if yes, it is judged that the pig has eaten, otherwise, it is judged that the pig has not eaten.

[0016] By adopting the technical scheme, the pig body abnormality or the damage of the feeder can be judged by judging that the pig has not eaten.

[0017] Preferably, the intelligent feeding system for pig farms further comprises an abnormality judgment module, which is configured to judge the pig body abnormality or the damage of the feeder when the pig has not eaten.

[0018] By adopting the technical scheme, the pig can be treated or processed when the pig body is abnormal, and the feeder can be repaired when the feeder is damaged.

[0019] Preferably, the abnormality judgment module comprises an infrared thermometer, which is configured to measure a body temperature of the pig, if the measured body temperature of the pig is abnormal, it is judged that the pig body is abnormal, otherwise, it is judged that the feeder is damaged.

[0020] By adopting the technical scheme, the pig body abnormality or the damage of the feeder can be effectively judged

[0021] Preferably, a back of the trough is arranged against a wall.

[0022] By adopting the technical scheme, the trough is arranged close to the wall, so that the trough is less likely to be damaged by the pigs, and the service life of the trough is prolonged. In addition, since there is a case that the pigs close to the back of the trough do not have a real feeding intention, the trough is arranged close to the wall, so that the pigs close to the back of the trough can be avoided, and the reliability of the feeding intention of the pigs is improved.

[0023] In a second aspect, the application further provides a control method of the pig farm intelligent feeding system, characterized in that the method comprises:

[0024] S1: acquiring pig video images and trough video images by using an image acquisition module;

[0025] S2: judging whether the pigs have a feeding intention by using a pig feeding intention judgment module, specifically comprising: recognizing pig skeleton joints in the pig video images by using a posture recognition algorithm, and judging whether a mouth joint of the pig is located within a preset trough peripheral distance and exceeds a preset time, yes, the pig has a feeding intention, otherwise, the pig does not have a feeding intention;

[0026] S3: in response to the judgment result of the pig feeding intention judgment module, controlling the feeding device according to the feeding intention of the pigs by using a control module.

[0027] The pig farm intelligent feeding system and the control method thereof disclosed by the application can automatically feed according to the feeding intention of the pigs, so as to meet the feeding needs of different individuals, improve the health of the pigs, and improve the breeding efficiency. The pig farm intelligent feeding system further comprises a control feeding amount module, the feeding amount of the pigs can be set according to the last feeding time of the pigs by the control feeding amount module, so that the feeding amount of the pigs can be optimized, which is beneficial to resource allocation. The pig farm intelligent feeding system further comprises an abnormality judgment module, which is used for judging whether the pigs have a body abnormality or the feeding device is damaged when the pigs do not feed, so that the pigs can be treated or processed when the pigs have a body abnormality, and the feeding device can be repaired when the feeding device is damaged. The trough is arranged close to the wall, so that the trough is less likely to be damaged by the pigs, and the service life of the trough is prolonged. In addition, since there is a case that the pigs close to the back of the trough do not have a real feeding intention, the trough is arranged close to the wall, so that the pigs close to the back of the trough can be avoided, and the reliability of the feeding intention of the pigs is improved. BRIEF DESCRIPTION OF DRAWINGS

[0028] The accompanying drawings are included to provide a further understanding of embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain principles of the present application. Other embodiments and many of the intended advantages of the present application will be readily appreciated as the same becomes better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale relative to each other. Like reference numerals designate corresponding similar parts.

[0029] Figure 1 is a schematic diagram of a pig farm intelligent feeding system based on posture recognition according to an embodiment of the present application.

[0030] Figure 2 is a piglet skeleton joint recognition schematic diagram.

[0031] Figure 3 is a flowchart of a control method of a pig farm intelligent feeding system according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0033] In the description of the present application, the terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0034] In a first aspect, the present application discloses a pig farm intelligent feeding system based on posture recognition, referring to Figure 1 The system specifically includes a feeder 100, an image acquisition module 110, a pig feeding willingness judgment module 120, and a control module 130.

[0035] The feeder 100 comprises a storage barrel 101, a dropping channel 102 and a trough 103. The feeder 100 is located above the trough 103. The dropping channel 102 connects the storage barrel 101 and the trough 103. The food in the storage barrel 101 falls into the trough 103 through the dropping channel 102. The trough 103 is used to receive the food dropped by the storage barrel 101 for the pigs to eat.

[0036] The image acquisition module 110 is used to acquire the video image of the pigs and the video image of the trough. In the embodiment, the image acquisition module 110 is a camera which is arranged on the top of the farm.

[0037] The pig eating intention judgment module 120 identifies the pig skeleton joints in the video image of the pigs by using a pose recognition algorithm, and judges whether the mouth joint of the pig is located within a preset distance from the periphery of the trough and exceeds a preset time. If yes, it is judged that the pig has the eating intention. If not, it is judged that the pig does not have the eating intention. Further, the pig eating intention judgment module 120 is also used to judge whether the mouth joint of the pig is blocked by the trough and exceeds a preset time. If yes, it is judged that the pig has the eating intention. If not, it is judged that the pig does not have the eating intention. Specifically, the judgment of whether the mouth joint of the pig is located within a preset distance from the periphery of the trough can be within a preset range with the trough as the center and a radius R, for example, R can be 10-50 cm; preferably, 10-30 cm, and in one of the embodiments, the preset range of R is within 20 cm. This is because animals generally have a certain stress. When they are used to eating at the trough, if they have a feeling of hunger or eating intention, they will generally be stressed and move around the trough. The distance between the mouth joint and the trough can be calculated by existing software, which is not repeated here. The preset time can be 20 seconds to 2 minutes, etc., which can be set according to actual needs. In one of the embodiments, the preset time is about 1 minute.

[0038] Reference Figure 2 In a specific embodiment, the pig eating intention judgment module 120 uses the OpenPose pose recognition algorithm to acquire all frames in the video segment and outputs the joints of the piglet skeleton in all frames, including the head, shoulder, left front elbow, right front elbow, hip, left rear elbow, right rear elbow and mouth joint, etc. The OpenPose pose recognition algorithm can predict the position of the partially blocked local joint. Further, the piglet pose can also be tracked by the PoseTrack pose tracking algorithm. The OpenPose pose recognition algorithm can identify up to 27 skeleton joints of the pig.

[0039] The control module 130 is in signal connection with the pig feeding intention judgment module 120, and the control valve 104 is arranged on the feeding channel 102 and in signal connection with the control module 130. In the embodiment, the control module 130 controls the control valve to be closed under normal circumstances or after the pig is fed, and the control module 130 controls the control valve to be opened when the pig has feeding intention, so as to control the feeder 100 to feed.

[0040] When the pig's "mouth closing point" is close to the trough within a predetermined distance (or blocks the trough), and exceeds the set time (for example, 1 minute), it is judged that the pig has feeding or watering intention to automatically feed. In the present application, the storage barrel can store food or water, and feeding and watering are generally two separate automatic feeding or watering systems.

[0041] In a further embodiment, the pig farm intelligent feeding system further comprises a control feeding amount module, which is used to set the feeding amount according to the last feeding time of the pig. Specifically, the amount of feeding can be optimized and configured in combination with the last feeding time. For example, if the last feeding time is lower than the preset time, a small amount of feeding (for example, the standard feeding of one meal is A kilograms, and the small amount of feeding is generally about 1 / 5A) is performed; otherwise, the standard feeding is performed, or the feeding is performed in other ways; (A is set according to different back fat conditions, in addition, the identification of back fat is prior art, which will not be described here).

[0042] In a further embodiment, the pig farm intelligent feeding system further comprises a feeding judgment module, which is used to judge whether the pig's mouth closing point coincides with or blocks the trough, and whether the coincidence or blocking exceeds the preset time, if so, it is judged that the pig has eaten, otherwise it is judged that the pig has not eaten.

[0043] In a further embodiment, the pig farm intelligent feeding system further comprises an abnormality judgment module, which is used to judge whether the pig's body is abnormal or the feeder is damaged when the pig does not eat. Specifically, if the pig does not eat (i.e., the mouth closing point does not coincide with or block the trough) after the automatic control of feeding is performed once, and the pig feeding intention judgment module 120 judges that the pig has multiple "feeding intention" (i.e., the pig's mouth closing point is within the preset distance of the trough periphery and exceeds the preset time) within a unit time; it can be judged that the pig's body is abnormal or the feeder is damaged and does not feed normally.

[0044] In some embodiments, the abnormality judging module comprises an infrared thermometer, which is used to measure the body temperature of the pig, if the body temperature of the pig is abnormal, it is judged that the pig is abnormal, and the pig is treated or other processing, otherwise it is judged that the feeder is damaged, and it is repaired. It can be understood that through the above judgment, whether the feeder is damaged or the pig's eating abnormal event is monitored can be determined without additional sensors, especially for sows in individual stalls.

[0045] In some embodiments, the abnormality judging module comprises a trough detection module, which is used to detect whether there is feed in the trough. Specifically, the video image of the trough can be trained by a neural network model to detect whether there is feed in the trough.

[0046] In further embodiments, the back of the trough is arranged against the wall. Arranging the trough against the wall makes the trough less likely to be damaged by the pig, prolonging the service life of the trough. In addition, since there are cases where the pig does not really have the intention to eat when it is close to the back of the trough, arranging the trough against the wall can avoid the pig being close to the back of the trough, thereby improving the reliability of the pig's eating intention.

[0047] In a second aspect, with reference to Figure 3 The application also discloses a control method of a pig farm intelligent feeding system based on posture recognition, which comprises:

[0048] S1: acquiring pig video images by using an image acquisition module and acquiring video images of the trough;

[0049] S2: judging whether the pig has an eating intention by using a pig eating intention judging module, specifically comprising: identifying the pig skeleton joint in the pig video image by using a posture recognition algorithm, and judging whether the pig's mouth joint is located within the preset trough peripheral distance and exceeds the preset time, if yes, it is judged that the pig has an eating intention, otherwise it does not have an eating intention;

[0050] S3: in response to the judgment result of the pig eating intention judging module, controlling the feeder according to the pig's eating intention by using a control module.

[0051] In specific embodiments, one specific embodiment of the control method of the pig farm intelligent feeding system of the application is disclosed as follows:

[0052] Step one: using OpenPose pose recognition algorithm, all frames in the video segment are obtained, and the piglet skeleton joints are output, including: head, shoulder, left front foot elbow, right front foot elbow, hip, left rear foot elbow, right rear foot elbow, and mouth joint, etc. OpenPose pose recognition algorithm can predict the position of the partially blocked local joint. Further, the piglet pose can be tracked by PoseTrack pose tracking algorithm. OpenPose pose recognition algorithm can identify up to 27 skeletal joints of live pigs.

[0053] Step two: automatic feeding cylinder and skeletal joint recognition linkage, when the "mouth joint" of the live pig is close to the trough within a predetermined distance (or blocks the trough), and exceeds the set time (for example, 1 minute), it is judged that the pig has the intention to eat or drink, and the automatic feeding is started.

[0054] Step three: the amount of feeding can be optimized in combination with the last feeding time. For example, if the last feeding time is less than the preset time, a small amount of feeding (for example, the standard feeding of one meal is A kilograms, and the small amount of feeding is generally about 1 / 5A) is performed; otherwise, the standard feeding is performed, or other ways of feeding are performed; (A is set according to different back fat conditions, in addition, the identification of back fat is prior art, which will not be described here).

[0055] Step four: whether the pig eats after the feeding (whether the mouth joint coincides with the trough or blocks the trough), whether the pig eats or whether the feeder is damaged. If the automatic control of the feeding is performed for several times and the live pig does not eat, that is, the mouth joint does not coincide with the trough or block the trough; or it can be judged that the live pig has a physical abnormality, or it can be judged that the feeder is damaged and does not feed normally.

[0056] In summary, the pig farm intelligent feeding system and the control method thereof based on pose recognition disclosed in the present application have the following beneficial technical effects:

[0057] 1. The pig farm intelligent feeding system and the control method thereof based on pose recognition disclosed in the present application can automatically feed according to the eating intention of the pig, thereby meeting the feeding needs of different individuals, improving the health of the pig, and improving the breeding efficiency;

[0058] 2. The pig farm intelligent feeding system further comprises a control feeding amount module, and the feeding amount module can set the feeding amount according to the last feeding time of the pig, thereby optimizing the feeding amount of the pig, and facilitating the resource allocation;

[0059] 3. The field intelligent feeding system further comprises an abnormality judging module, the abnormality judging module is used for judging whether the pig is abnormal or the feeder is damaged when the pig does not eat, so that the pig can be treated or other processed when the pig is abnormal, and the feeder can be repaired when the feeder is damaged.

[0060] 4. The trough is arranged close to the wall, so that the trough is not easily damaged by the pig, and the service life of the trough is prolonged. In addition, since there is a case that the pig close to the back of the trough does not really have the intention to eat, the trough is arranged close to the wall, so that the pig close to the back of the trough is avoided, thereby improving the reliability of the intention of the pig to eat.

[0061] The application further provides a computer readable storage medium, which stores a program file, and the program file is used for executing the method.

[0062] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product implemented on one or more computer readable storage media containing computer usable program code (including but not limited to disk memory, CD-ROM, optical memory, etc.).

[0063] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system) and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be realized by computer program instructions. These computer program instructions can be provided to a general purpose computer, a special purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the computer or other programmable data processing device produce a device for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The means for implementing the functions specified in one or more flows and / or blocks.

[0064] These computer program instructions can also be stored in a computer readable storage medium, which can guide the computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer readable storage medium produce a product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The means for implementing the functions specified in one or more flows and / or blocks.

[0065] These computer program instructions can also be loaded into a computer or other programmable data processing devices, so that a series of operational steps are performed on the computer or other programmable data processing devices to generate a computer implemented process, so that the instructions executed on the computer or other programmable data processing devices provide a process for implementing the functions specified in the flowchart Figure 1 one flow or a plurality of flows and / or the functions specified in the block Figure 1 one flow or a plurality of flows and / or the functions specified in the block

[0066] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A pig farm intelligent feeding system based on posture recognition, characterized in that, The pig farm intelligent feeding system comprises the following: a feeder comprising a storage barrel, a material falling channel connecting the storage barrel and a trough, and the trough being used for receiving food falling from the storage barrel for pigs to eat; an image acquisition module used for acquiring a video image of the pig and a video image of the trough; a pig eating intention judgment module used for identifying a pig skeleton joint in the pig video image by using a posture recognition algorithm, judging whether a mouth joint of the pig is located within a preset distance from the periphery of the trough and exceeds a preset time, and judging that the pig has an eating intention if yes, and otherwise judging that the pig has no eating intention; the judgment of whether the mouth joint of the pig is located within the preset distance from the periphery of the trough is within a preset range with the trough as a center and a radius R, and R is 10-30 cm; the preset time is 20 seconds-2 minutes; a control module in signal connection with the pig eating intention judgment module, and a control valve is arranged on the material falling channel and in signal connection with the control module; the pig farm intelligent feeding system further comprises a control feeding amount module used for setting a feeding amount according to a last feeding time of the pig; specifically, if the last feeding time of the pig is lower than a preset time, a small amount of feeding is performed, otherwise, standard feeding or overfeeding is performed; the standard feeding of one meal is A kilograms, and the small amount of feeding is 1 / 5A; the pig farm intelligent feeding system further comprises an eating judgment module used for judging whether the mouth joint of the pig coincides with or is shielded by the trough, and whether the coincidence or shielding exceeds a preset time, and judging that the pig has eaten if yes, and otherwise judging that the pig has not eaten; the pig farm intelligent feeding system further comprises an abnormality judgment module used for judging whether the pig is abnormal or the feeder is damaged when the pig has not eaten; the abnormality judgment module comprises an infrared thermometer used for measuring a body temperature of the pig, and judging that the pig is abnormal if the measured body temperature of the pig is abnormal, and otherwise judging that the feeder is damaged; the abnormality judgment module further comprises a trough detection module used for detecting whether there is material in the trough.

2. The pig farm intelligent feeding system based on gesture recognition according to claim 1, characterized in that: The back of the trough is arranged against a wall.

3. The control method of the pig farm intelligent feeding system based on gesture recognition according to claim 1, characterized in that, The method comprises the following steps: S1: acquiring a video image of the pig and a video image of the trough by using an image acquisition module; S2: judging whether the pig has an eating intention by using a pig eating intention judgment module, specifically comprising: identifying a pig skeleton joint in the pig video image by using a posture recognition algorithm, judging whether a mouth joint of the pig is located within a preset distance from the periphery of the trough and exceeds a preset time, and judging that the pig has an eating intention if yes, and otherwise judging that the pig has no eating intention; S3: in response to a judgment result of the pig eating intention judgment module, controlling the feeder to fall material according to the eating intention of the pig by using a control module.

Images