Ruminant detection method and device of ruminant, electronic equipment and storage medium

Abstract

The invention provides a ruminant detection method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the posture recognition of a ruminant, and obtaining a target posture; the target posture comprises a standing posture state and a prone posture state; acquiring an audio signal of the ruminant animal in the target posture; inputting the audio signal into a rumination signal identification model to obtain a rumination state prediction sequence; under the condition that the signal corresponding to the current state in the ruminant state prediction sequence is a ruminant inverse vomiting signal, acquiring echo signals obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signals; and determining rumination duration based on the moments corresponding to the enhanced echo signals. Rumination actions are actively detected by using the patch type ultrasonic sensor, so that the problem that rumination is easily infected by noise by generally detecting the rumination purely depending on sound is avoided; the rumination process of the ruminant is accurately monitored by using the rumination signal recognition model, and the method has the advantages of high accuracy, strong real-time performance and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of rumination detection, and in particular to a rumination detection method, device, electronic equipment and storage medium for ruminants. Background Art

[0002] Rumination is a unique way of digestion for ruminants such as cattle, which is closely related to the function of the forestomach and abomasum of cattle and their overall health status. By observing the rumination activities of cattle, we can intuitively understand the health status of cattle. Regularly testing the rumination activities of cattle can timely discover digestive problems of cattle, so as to take corresponding preventive measures to avoid the occurrence of diseases. It can also provide an important basis for feeding management, formulate more scientific and reasonable feeding plans and management measures, and improve the economic benefits of animal husbandry.

[0003] Currently, the methods for detecting cattle rumination mainly include rumination sound detection, behavior detection, visual detection, neck deformation detection and rumen pressure detection.

[0004] Existing detection based on neck deformation or nasal cavity detection may cause great discomfort to cattle and may even affect their health. In addition to the discomfort that may be caused, implantable devices such as rumen capsules also have problems such as limited lifespan and difficulty in replacement. Visual detection requires accurate identification and tracking of cattle and precise focus and magnification of local areas. However, there are many cattle on cattle farms and it is very easy to lose the target. A more practical method is rumination sound detection, but each cow has different breathing, calling, rumination, and swallowing habits, and the sound is relatively weak. The environmental noise is high, and it is difficult for a generalized model to achieve accurate detection, and the false alarm rate is high. Summary of the invention

[0005] The present invention provides a rumination detection method, device, electronic device and storage medium for ruminants, so as to solve the problem that neck deformation detection or nasal cavity detection in the prior art may cause great discomfort to cattle and may even affect their health. In addition to the possibility of causing discomfort, implantable devices such as rumen capsules also have the defects of limited lifespan and difficulty in replacement.

[0006] The present invention provides a rumination detection method for ruminants, comprising the following steps: Performing posture recognition on the ruminant to obtain a target posture; the target posture includes a standing posture state and a lying posture state; Acquiring an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model; When the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, acquiring an echo signal obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signal; The rumination duration of the ruminant is determined based on the time instants corresponding to the enhanced echo signals.

[0007] According to a rumination detection method for ruminants provided by the present invention, the rumination state prediction sequence includes a first state in a standing state and a second state in a lying state, and the first state and the second state both include at least one of a rumination and vomiting state, a swallowing state, a feeding and chewing state, a rumination and chewing state, and no action.

[0008] According to a ruminant rumination detection method provided by the present invention, the training step of the ruminant signal recognition model comprises: Acquire an initial rumination signal recognition model, and acquire a first sample audio signal in a standing state, and a first label state of the first sample audio signal; Inputting the first sample audio signal into an initial rumination signal recognition model to obtain a first rumination state prediction result output by the initial rumination signal recognition model; determining a first loss based on a difference between the first rumination state prediction result and the first label state, and performing parameter iteration on the initial rumination signal recognition model based on the first loss to obtain a candidate rumination signal recognition model; Acquire a second sample audio signal in a lying state, and a second label state of the second sample audio signal; Inputting the second sample audio signal into the candidate rumination signal recognition model to obtain a second rumination state prediction result output by the candidate rumination signal recognition model; Based on the difference between the second rumination state prediction result and the second label state, a second loss is determined, and based on the second loss, parameters of the candidate rumination signal recognition model are iterated to obtain the rumination signal recognition model.

[0009] According to a ruminant rumination detection method provided by the present invention, the rumination duration of the ruminant is determined based on the time corresponding to each enhanced echo signal, including: Determine the first moment when the signal corresponding to the current state is a rumination and vomiting signal, and use the first moment as the rumination start moment of the ruminant; Determining the end time of each enhanced echo signal, and when none of the enhanced echo signals is detected within a preset time period, taking the end time of the last enhanced echo signal among the enhanced echo signals as the end time of rumination of the ruminant; The rumination duration is determined based on the rumination start time and the rumination end time.

[0010] The present invention also provides a rumination detection device for ruminants, comprising an ultrasonic sensor, a sound pressure sensor, a rumination monitor and a low-power posture monitor; The ultrasonic sensor is fixed behind the mandibular gland of the ruminant and directly facing the esophagus, and is used to collect the ultrasonic signal of the ruminant; The low-power posture monitor is installed on the foot of the ruminant to detect the target posture of the ruminant; The sound pressure sensor is fixed to the upper part of the cervical fasciitis muscle behind the ear of the ruminant, and is used to collect the audio signal of the ruminant in the target posture; The rumination monitor is used to input the audio signal into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model, and when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain the echo signal obtained by performing ultrasonic signal detection on the ruminant, determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0011] According to a ruminant rumination detection device provided by the present invention, the low-power posture monitor includes a gyroscope, a first single-chip microcomputer, a first low-power wireless communication module, a power management chip and an indicator light; The gyroscope is used to detect the target posture of the ruminant, obtain posture information, and upload the posture information to the single-chip microcomputer; The first single-chip microcomputer receives the posture information, and uploads the posture information to the rumination monitor based on the first low-power wireless communication module; The indicator light is used to indicate the working status of the single chip microcomputer.

[0012] According to a rumination detection device for ruminants provided by the present invention, the rumination monitor includes a second single-chip microcomputer, a 485 communication module and a second low-power wireless communication module; The 485 communication module is used to transmit the ultrasonic signal collected by the ultrasonic sensor to the second single-chip microcomputer, and the 485 communication module is also used to transmit the audio signal collected by the sound pressure sensor to the second single-chip microcomputer; The second low-power wireless communication module performs wireless communication with the first low-power wireless communication module; The second single-chip computer is used to input the audio signal into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model, and when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain an echo signal obtained by ultrasonic signal detection of the ruminant, and determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal. The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a ruminant detection method as described above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the rumination detection method for a ruminant as described in any one of the above is implemented.

[0014] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for detecting rumination of ruminants.

[0015] The rumination detection method, device, electronic device and storage medium of ruminants provided by the present invention perform posture recognition on ruminants to obtain target posture; the target posture includes standing posture and lying posture; obtain the audio signal of the ruminant in the target posture; input the audio signal into the rumination signal recognition model to obtain the rumination state prediction sequence output by the rumination signal recognition model; when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain the echo signal obtained by ultrasonic signal detection of the ruminant, and determine each enhanced echo signal in the echo signal; based on the time corresponding to each enhanced echo signal, determine the rumination duration of the ruminant. On the one hand, the patch type ultrasonic sensor is used to actively detect the rumination action, avoiding the problem that the system that generally relies solely on sound to detect rumination is susceptible to noise infection; on the other hand, the rumination signal recognition model is used to realize the accurate monitoring of the rumination process of the ruminant and the accurate calculation of the rumination amount. The method has the advantages of simple operation, high accuracy, strong real-time performance, etc., and provides strong support for the health management and production performance evaluation of ruminants. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram of the process of the rumination detection method of ruminants provided by the present invention.

[0018] Figure 2 It is a schematic diagram of the positions of the ultrasonic sensor and the sound pressure sensor provided by the present invention.

[0019] Figure 3 It is a schematic structural diagram of a ruminant rumination detection device provided by the present invention.

[0020] Figure 4 It is a schematic diagram of the low-power attitude monitor provided by the present invention.

[0021] Figure 5 Schematic diagram of a collar provided by the present invention.

[0022] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, rather than to describe a specific order or precedence. It should be understood that the terms used in this way can be interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type.

[0025] Figure 1 : is a schematic diagram of the process of the ruminant rumination detection method provided by the present invention, such as Figure 1 As shown, the method includes step 110 , step 120 , step 130 and step 140 .

[0026] Step 110, performing posture recognition on the ruminant to obtain a target posture; the target posture includes a standing posture state and a lying posture state.

[0027] Specifically, the embodiment of the present invention takes into account that the audio signals of ruminants in different postures are different, therefore, firstly, posture recognition is performed on the ruminants to identify whether the ruminants are in a standing state or a lying state. That is, the target posture includes a standing state and a lying state.

[0028] Here, the posture recognition of the ruminant can be achieved by a low-power posture monitor, which is installed on the foot of the ruminant to detect the target posture of the ruminant.

[0029] Here, the ruminant may be a cow, a sheep, a camel, etc., and this embodiment of the present invention does not specifically limit this.

[0030] Step 120: acquiring an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model.

[0031] Specifically, an audio signal of the ruminant in the target posture may be obtained. Here, the audio signal may be obtained by a patch-type sound pressure sensor, which is fixed to the upper part of the cervical fasciitis muscle of the neck behind the ear of the ruminant.

[0032] It is understandable that when ruminants are eating, ruminating or swallowing, their oral and throat areas will emit audio signals with certain characteristics, and these audio signals can be used to reflect the current rumination state.

[0033] Here, the audio signal of the ruminant may be acquired in real time or at intervals, which is not specifically limited in the embodiment of the present invention.

[0034] It should be noted that after the audio signal of the ruminant is obtained, the audio signal may be filtered, for example, the audio signal in the range of 200 Hz-2000 Hz may be extracted after filtering.

[0035] After the audio signal is obtained, the audio signal may be input into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model.

[0036] Here, the rumination signal recognition model may be a long short-term memory network (Long Short-Term Memory, LSTM) or the like, and the embodiment of the present invention does not specifically limit this.

[0037] Among them, the rumination state prediction sequence includes a first state in a standing state and a second state in a lying state, and the first state and the second state both include at least one of a rumination and vomiting state, a swallowing state, a feeding and chewing state, a rumination and chewing state, and no action.

[0038] It is understandable that the embodiment of the present invention takes into account that the audio signals in the standing state and the lying state are different, therefore, the audio signal is input into the rumination signal recognition model to obtain the rumination state prediction sequence output by the rumination signal recognition model. In addition, the rumination state prediction sequence is used to reflect that the state corresponding to the current signal is the rumination and vomiting state, swallowing state, feeding and chewing state, rumination and chewing state, and no action in the first state or the second state.

[0039] It should be noted that the embodiment of the present invention utilizes the rumination signal recognition model to process the audio signal and the posture timing signal to accurately identify 10 states of rumination and vomiting, swallowing, eating and chewing, rumination and chewing action and no action in standing and lying states, and can also reduce the disturbance of posture to the rumination audio.

[0040] Step 130, when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, an echo signal obtained by performing ultrasonic signal detection on the ruminant is obtained, and each enhanced echo signal in the echo signal is determined.

[0041] Specifically, when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, an echo signal obtained by ultrasonic signal detection of the ruminant is obtained. Here, ultrasonic signal detection of the ruminant can be achieved by a patch type ultrasonic sensor. Figure 2 Schematic diagram of the position of the ultrasonic sensor and the sound pressure sensor provided by the present invention, such as Figure 2 As shown, the patch-type ultrasonic sensor is fixed behind the mandibular gland of the ruminant and directly opposite the esophagus.

[0042] For example, when a rumination or vomiting signal is detected by a patch-type sound pressure sensor, this moment is marked as the start of rumination, and the patch-type ultrasonic sensor is started. The patch-type ultrasonic sensor detects the echo signal at a frequency of 10 Hz and determines each enhanced echo signal in the echo signal, that is, the rumination or vomiting state is accurately identified according to the frequency and amplitude of the signal.

[0043] It can be understood that by utilizing the enhanced reflection characteristics of ultrasound on air and soft tissue and using patch-type ultrasonic sensors to actively detect rumination actions, the problem of the general system that relies solely on sound to detect rumination being susceptible to noise infection can be avoided, thereby further improving the accuracy and reliability of subsequent rumination detection.

[0044] Step 140: Determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0045] Specifically, the end time of each enhanced echo signal is determined, and based on the end time corresponding to each enhanced echo signal, the rumination duration of the ruminant is determined.

[0046] For example, the number of enhanced echo signals is counted, and if continuous esophageal movement is detected (i.e., continuous enhanced echo signals), the rumination amount is calculated according to the number of enhanced echo signals. If no continuous esophageal movement is detected, the rumination and vomiting are marked as invalid.

[0047] If no enhanced echo signal is detected for a preset period of time, the moment of the last enhanced echo signal is counted as the end time of rumination, and then the first moment when the signal corresponding to the current state is a rumination and vomiting signal is determined as the start time of rumination. Based on the start time and the end time of rumination, the rumination duration of the ruminant is determined.

[0048] The method provided by the embodiment of the present invention performs posture recognition on ruminants to obtain target postures; the target postures include standing and lying; obtains the audio signal of the ruminant in the target posture; inputs the audio signal into the ruminant signal recognition model to obtain the ruminant state prediction sequence output by the ruminant signal recognition model; when the signal corresponding to the current state in the ruminant state prediction sequence is a ruminant vomiting signal, obtains the echo signal obtained by ultrasonic signal detection of the ruminant, and determines each enhanced echo signal in the echo signal; based on the time corresponding to each enhanced echo signal, determines the rumination duration of the ruminant. On the one hand, a patch-type ultrasonic sensor is used to actively detect the rumination action, avoiding the problem that the general system that relies solely on sound detection of rumination is susceptible to noise infection; on the other hand, the ruminant signal recognition model is used to realize the accurate monitoring of the ruminant rumination process and the accurate calculation of the rumination amount. The method has the advantages of simple operation, high accuracy, strong real-time performance, etc., and provides strong support for the health management and production performance evaluation of ruminants.

[0049] Based on the above embodiment, the training step of the rumination signal recognition model includes: Step 210, obtaining an initial rumination signal recognition model, and obtaining a first sample audio signal in a standing state, and a first label state of the first sample audio signal; Step 220, inputting the first sample audio signal into an initial ruminant signal recognition model to obtain a first ruminant state prediction result output by the initial ruminant signal recognition model; Step 230, determining a first loss based on the difference between the first rumination state prediction result and the first label state, and performing parameter iteration on the initial rumination signal recognition model based on the first loss to obtain a candidate rumination signal recognition model; Step 240, obtaining a second sample audio signal in a lying state, and a second label state of the second sample audio signal; Step 250, inputting the second sample audio signal into the candidate rumination signal recognition model to obtain a second rumination state prediction result output by the candidate rumination signal recognition model; Step 260: determine a second loss based on the difference between the second rumination state prediction result and the second label state, and perform parameter iteration on the candidate rumination signal recognition model based on the second loss to obtain a rumination signal recognition model.

[0050] Specifically, in order to better obtain the rumination signal recognition model, training can be performed based on the following steps: First, an initial rumination signal recognition model is obtained, and a first sample audio signal in a standing state and a first label state of the first sample audio signal are obtained.

[0051] The parameters of the initial rumination signal recognition model may be preset or randomly generated, and the initial rumination signal recognition model may be an LSTM model, which is not specifically limited in the embodiment of the present invention. Here, the first label state includes the rumination and vomiting state, the swallowing state, the feeding and chewing state, the rumination and chewing state, and no action in the standing state.

[0052] It should be noted that the standing state and the lying state can be detected by a low-power posture monitor installed on the feet of the ruminant.

[0053] Then, the first sample audio signal may be input into the initial ruminant signal recognition model to obtain a first ruminant state prediction result output by the initial ruminant signal recognition model.

[0054] After obtaining the first ruminative state prediction result based on the initial ruminative signal recognition model, the first loss can be determined based on the difference between the first ruminative state prediction result and the first label state, and the parameters of the initial ruminative signal recognition model can be iterated based on the first loss. The initial ruminative signal recognition model after the parameter iteration is used as a candidate ruminative signal recognition model.

[0055] It can be understood that the greater the difference between the first rumination state prediction result and the first label state, the greater the first loss; the smaller the difference between the first rumination state prediction result and the first label state, the smaller the first loss.

[0056] After obtaining the candidate rumination signal recognition model, the rumination signal recognition training in the lying state can be carried out as follows: First, a second sample audio signal in a lying state and a second label state of the second sample audio signal are obtained. Here, the second label state includes a rumination and vomiting state, a swallowing state, a feeding and chewing state, a rumination and chewing state, and no action in a standing state.

[0057] Then, the second sample audio signal is input into the candidate ruminant signal recognition model to obtain a second ruminant state prediction result output by the candidate ruminant signal recognition model.

[0058] After obtaining the second ruminative state prediction result based on the candidate ruminative signal recognition model, the second loss can be determined based on the difference between the second ruminative state prediction result and the second label state, and the parameters of the candidate ruminative signal recognition model can be iterated based on the second loss. The candidate ruminative signal recognition model after the parameter iteration is used as the ruminative signal recognition model.

[0059] It can be understood that the greater the difference between the second rumination state prediction result and the second label state, the greater the second loss; the smaller the difference between the second rumination state prediction result and the second label state, the smaller the second loss.

[0060] The method provided by the embodiment of the present invention combines the low-power posture monitor information, audio signal and ultrasonic information, and the processor automatically collects the sound pressure and ultrasonic data in the lying state and the standing state, and self-learns and optimizes the ruminant signal recognition model, thereby improving the recognition accuracy and reliability of the ruminant signal recognition model.

[0061] Based on the above embodiment, step 140 includes: Step 141, determining the first moment when the signal corresponding to the current state is a rumination and vomiting signal, and taking the first moment as the rumination start moment of the ruminant; Step 142, determining the end time of each enhanced echo signal, and when the enhanced echo signals are not detected within a preset time period, taking the end time of the last enhanced echo signal among the enhanced echo signals as the end time of rumination of the ruminant; Step 143: Determine the rumination duration based on the rumination start time and the rumination end time.

[0062] Specifically, it is determined that the signal corresponding to the current state is the first moment of the rumination and vomiting signal, and the first moment is used as the rumination start time of the ruminant.

[0063] Then, the end time of each enhanced echo signal is determined, and when no enhanced echo signal is detected within a preset time period, the end time of the last enhanced echo signal among the enhanced echo signals is used as the end time of rumination of the ruminant. The last enhanced echo signal is the last enhanced echo signal.

[0064] Here, the preset duration may be 2 minutes, or 3 minutes, etc., and this is not specifically limited in the embodiment of the present invention.

[0065] Finally, based on the start and end times of rumination, the duration of rumination can be determined, and the amount of rumination can also be counted.

[0066] The rumination detection device for ruminants provided by the present invention is described below. The rumination detection device for ruminants described below and the rumination detection method for ruminants described above can be referenced to each other.

[0067] Based on any of the above embodiments, the present invention provides a rumination detection device for ruminants, Figure 3 Schematic diagram of the structure of the ruminant detection device provided by the present invention, such as Figure 3 As shown, the rumination detection device includes an ultrasonic sensor 10, a sound pressure sensor 20, a rumination monitor 30 and a low-power posture monitor 40; The ultrasonic sensor 10 is fixed behind the mandibular gland of the ruminant and facing the esophagus, and is used to collect the ultrasonic signal of the ruminant; The low-power posture monitor 40 is installed on the foot of the ruminant to detect the target posture of the ruminant; The sound pressure sensor 20 is fixed to the upper part of the cervical fasciitis muscle behind the ear of the ruminant, and is used to collect the audio signal of the ruminant in the target posture; The rumination monitor 30 is used to input the audio signal into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model, and when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain the echo signal obtained by performing ultrasonic signal detection on the ruminant, determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0068] The rumination monitor 30 includes a second single chip microcomputer, a 485 communication module and a second low power wireless communication module (LORA, Long Range).

[0069] The 485 communication module is used to transmit the ultrasonic signal collected by the ultrasonic sensor 10 to the second single chip microcomputer, and the 485 communication module is also used to transmit the audio signal collected by the sound pressure sensor 20 to the second single chip microcomputer.

[0070] The second low-power wireless communication module performs wireless communication with the first low-power wireless communication module.

[0071] The second single-chip computer is used to input the audio signal into the ruminant signal recognition model to obtain the ruminant state prediction sequence output by the ruminant signal recognition model, and when the signal corresponding to the current state in the ruminant state prediction sequence is a ruminant vomiting signal, obtain the echo signal obtained by ultrasonic signal detection of the ruminant, determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0072] The ultrasonic sensor 10 and the rumination monitor 30 communicate with each other via a 485 communication module. RS485 refers to a serial physical interface standard established by the Electronic Industry Association (EIA) of the United States.

[0073] In addition, the rumination monitor 30 also includes a network communication serial port UART (Universal Asynchronous Receiver / Transmitter).

[0074] The rumination monitor 30 is provided with a single-chip microcomputer, which is connected to a LORA (Long Range) module. The LORA module is a low-power wireless communication module. The LoRa module is designed to achieve long-distance, low-power data transmission and is very suitable for Internet of Things applications.

[0075] The main features of the LoRa module include: Low power consumption: LoRa modules are designed to run on battery-powered devices for a long time, which can significantly extend the service life of the device. This is due to its efficient low-power design, which makes LoRa modules have significant energy-saving advantages in IoT applications.

[0076] Long-distance communication: Thanks to its efficient spread spectrum modulation technology, the LoRa module can achieve a communication distance of several kilometers in urban environments, and even tens of kilometers in suburban or rural areas. This makes the LoRa module have a wide range of application prospects in IoT applications that require long-distance communication.

[0077] In addition, the ruminant monitor 30 further includes an indicator light for indicating the working state of the single-chip microcomputer in the ruminant monitor. The ruminant monitor 30 further includes a power conversion circuit for performing power conversion to charge the single-chip microcomputer. The ruminant detection device for ruminants further includes an external power supply for charging the ruminant monitor 30.

[0078] In addition, the sensor can be tightly attached to the skin of the ruminant using a collar and appropriate weights to ensure that the sensor can stably collect signals.

[0079] The device provided by the embodiment of the present invention performs posture recognition on ruminants to obtain target postures; the target postures include standing and lying; the audio signal under the target posture is input into the rumination signal recognition model to obtain the rumination state prediction sequence output by the rumination signal recognition model; when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, the echo signal obtained by ultrasonic signal detection of the ruminant is obtained, and each enhanced echo signal in the echo signal is determined; based on the time corresponding to each enhanced echo signal, the rumination duration of the ruminant is determined. On the one hand, the patch-type ultrasonic sensor is used to actively detect the rumination action, avoiding the problem that the general system that relies solely on sound detection of rumination is susceptible to noise infection; on the other hand, the rumination signal recognition model is used to realize the accurate monitoring of the rumination process of the ruminant and the accurate calculation of the rumination amount. The method has the advantages of simple operation, high accuracy, strong real-time performance, etc., and provides strong support for the health management and production performance evaluation of ruminants.

[0080] It should be noted that the low-power posture monitor 40 can be set on a foot ring installed on the foot of the ruminant to detect the standing and lying posture changes of the ruminant, and report the posture change information to the ruminant monitor through the low-power Bluetooth communication module.

[0081] Figure 4 is a schematic diagram of a low power consumption attitude monitor provided by the present invention, such as Figure 4 As shown, the low power posture monitor 40 includes a gyroscope, a first single chip microcomputer, a first low power wireless communication module, a power management chip and an indicator light.

[0082] The gyroscope is used to detect the target posture of the ruminant, obtain the posture information, and upload the posture information to the microcontroller.

[0083] The first single chip microcomputer receives the posture information and uploads the posture information to the rumination monitor based on the first low power wireless communication module. The indicator light is used to indicate the working state of the single chip microcomputer.

[0084] The first single chip microcomputer in the low power posture monitor 40 is connected to the LORA module (first low power wireless communication module), and the LORA module is a low power wireless communication module.

[0085] In addition, the low power consumption attitude monitor 40 is also provided with a power management chip and a built-in battery. The built-in battery is used to charge the single chip microcomputer.

[0086] Based on any of the above embodiments, an embodiment of the present invention provides a method for detecting rumination in a dairy cow, comprising: fixing a patch-type sound pressure sensor on the upper part of the occipital muscle of the cow's neck behind its ear, fixing a patch-type ultrasonic sensor on the position of the esophagus behind the mandibular gland, Figure 5Schematic diagram of the collar provided by the present invention, such as Figure 5 As shown, use the collar and weight to hold the sensor against your skin. Figure 5 The top one is the sound pressure sensor, the right side of the middle part is the ultrasonic sensor, the left side is the counterweight for balancing the ultrasonic sensor, the bottom one is the total counterweight, and the main control unit is set to collect sensor signals; at the same time, foot rings are installed on the cow's feet to detect the cow's standing and lying posture.

[0087] After the foot ring detects the posture change, it reports it to the main control unit through the low-power Bluetooth communication module.

[0088] The main control unit detects the sound pressure signal every second, extracts the audio signal in the 200Hz-2000Hz range after filtering, trains the long short-term memory network LSTM, and recognizes rumination, vomiting, swallowing, eating and chewing, rumination and chewing movements and other 10 states in standing and lying states.

[0089] During the rumination process of dairy cows, the esophagus will be expanded and the pharyngeal muscles will be tightened due to vomiting and swallowing, and the ultrasonic echo signal intensity will increase significantly. After the main control unit detects the rumination and vomiting signal through the sound pressure sensor, this moment is marked as the start time of rumination, and the ultrasonic sensor is started to detect the echo signal at 10Hz. The rumination and vomiting state is accurately identified according to the signal frequency and amplitude, and the number of enhanced echo signals is counted. If a continuous echo enhancement signal is detected, the rumination amount is calculated according to the number of enhanced echo signals; if no continuous esophageal movement is detected, the rumination and vomiting is marked as invalid. After no echo enhancement is detected for 2 minutes, the last echo enhancement signal is counted as the end time of rumination.

[0090] Figure 6 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 6As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the rumination detection method of the ruminant, the method comprising: performing posture recognition on the ruminant to obtain a target posture; the target posture includes a standing state and a lying state; obtaining an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model; when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtaining an echo signal obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signal; and determining the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0091] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0092] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the rumination detection method for ruminants provided by the above-mentioned methods, the method including: performing posture recognition on the ruminant to obtain a target posture; the target posture includes a standing state and a lying state; obtaining an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model; when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtaining an echo signal obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signal; based on the time corresponding to each enhanced echo signal, determining the rumination duration of the ruminant.

[0093] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the rumination detection method for ruminants provided by the above-mentioned methods, the method comprising: performing posture recognition on the ruminant to obtain a target posture; the target posture comprises a standing state and a lying state; obtaining an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model; when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtaining an echo signal obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signal; and determining the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

[0094] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0095] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting rumination of ruminants, characterized in that: include: Perform posture recognition on ruminants to obtain target posture; The target posture includes a standing state and a lying state; Acquiring an audio signal of the ruminant in the target posture, inputting the audio signal into a rumination signal recognition model, and obtaining a rumination state prediction sequence output by the rumination signal recognition model; When the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, acquiring an echo signal obtained by performing ultrasonic signal detection on the ruminant, and determining each enhanced echo signal in the echo signal; The rumination duration of the ruminant is determined based on the time instants corresponding to the enhanced echo signals.

2. The ruminant rumination detection method according to claim 1, characterized in that: The rumination state prediction sequence includes a first state in a standing state and a second state in a lying state, and the first state and the second state both include at least one of a rumination and vomiting state, a swallowing state, a feeding and chewing state, a rumination and chewing state, and no action.

3. The ruminant rumination detection method according to claim 1, characterized in that: The training steps of the rumination signal recognition model include: Acquire an initial rumination signal recognition model, and acquire a first sample audio signal in a standing state, and a first label state of the first sample audio signal; Inputting the first sample audio signal into an initial rumination signal recognition model to obtain a first rumination state prediction result output by the initial rumination signal recognition model; determining a first loss based on a difference between the first rumination state prediction result and the first label state, and performing parameter iteration on the initial rumination signal recognition model based on the first loss to obtain a candidate rumination signal recognition model; Acquire a second sample audio signal in a lying state, and a second label state of the second sample audio signal; Inputting the second sample audio signal into the candidate rumination signal recognition model to obtain a second rumination state prediction result output by the candidate rumination signal recognition model; Based on the difference between the second rumination state prediction result and the second label state, a second loss is determined, and based on the second loss, parameters of the candidate rumination signal recognition model are iterated to obtain the rumination signal recognition model.

4. The method for detecting rumination of a ruminant according to any one of claims 1 to 3, characterized in that: The determining the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal comprises: Determine the first moment when the signal corresponding to the current state is a rumination and vomiting signal, and use the first moment as the rumination start moment of the ruminant; Determining the end time of each enhanced echo signal, and when none of the enhanced echo signals is detected within a preset time period, taking the end time of the last enhanced echo signal among the enhanced echo signals as the end time of rumination of the ruminant; The rumination duration is determined based on the rumination start time and the rumination end time.

5. A rumination detection device for ruminants, characterized in that: These include ultrasonic sensors, sound pressure sensors, rumination monitors, and low-power attitude monitors; The ultrasonic sensor is fixed behind the mandibular gland of the ruminant and directly facing the esophagus, and is used to collect the ultrasonic signal of the ruminant; The low-power posture monitor is installed on the foot of the ruminant to detect the target posture of the ruminant; The sound pressure sensor is fixed to the upper part of the cervical fasciitis muscle behind the ear of the ruminant, and is used to collect the audio signal of the ruminant in the target posture; The rumination monitor is used to input the audio signal into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model, and when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain the echo signal obtained by performing ultrasonic signal detection on the ruminant, determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

6. The ruminant rumination detection device according to claim 5, characterized in that: The low-power attitude monitor includes a gyroscope, a first single-chip microcomputer, a first low-power wireless communication module, a power management chip and an indicator light; The gyroscope is used to detect the target posture of the ruminant, obtain posture information, and upload the posture information to the single-chip microcomputer; The first single-chip microcomputer receives the posture information, and uploads the posture information to the rumination monitor based on the first low-power wireless communication module; The indicator light is used to indicate the working status of the single chip microcomputer.

7. The ruminant rumination detection device according to claim 5, characterized in that: The rumination monitor includes a second single chip microcomputer, a 485 communication module and a second low-power wireless communication module; The 485 communication module is used to transmit the ultrasonic signal collected by the ultrasonic sensor to the second single-chip microcomputer, and the 485 communication module is also used to transmit the audio signal collected by the sound pressure sensor to the second single-chip microcomputer; The second low-power wireless communication module performs wireless communication with the first low-power wireless communication module; The second single-chip microcomputer is used to input the audio signal into a rumination signal recognition model to obtain a rumination state prediction sequence output by the rumination signal recognition model, and when the signal corresponding to the current state in the rumination state prediction sequence is a rumination and vomiting signal, obtain an echo signal obtained by performing ultrasonic signal detection on the ruminant, determine each enhanced echo signal in the echo signal, and determine the rumination duration of the ruminant based on the time corresponding to each enhanced echo signal.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the rumination detection method for a ruminant according to any one of claims 1 to 4 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the rumination detection method for ruminants according to any one of claims 1 to 4 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the rumination detection method for ruminants according to any one of claims 1 to 4 is implemented.

Images