Pig injection intelligent identification method

Through laser beam measurement and CNN model recognition technology, the problems of pig skin thickness detection and injection parameter adjustment are solved, and the high accuracy and uniformity of pig injections are achieved, and the immune effect is improved.

CN118936308APending Publication Date: 2024-11-12HEFEI LASSETER ROBOT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411129241.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The prior art cannot accurately detect the pig skin thickness in the pig injection area, resulting in uneven vaccine injection or insufficient injection depth, affecting the immune effect.

Method used

By irradiating the pig surface with a laser beam, measuring the round-trip time and angle changes of the laser beam, calculating the three-dimensional coordinate information of the pig surface, forming three-dimensional point cloud data, and identifying the injection area through point cloud segmentation algorithm and CNN neural network model, and dynamically adjusting Injection parameters of the syringe to match the pig skin thickness.

Benefits of technology

High-precision measurement of pig skin thickness and precise adjustment of injection parameters are achieved, ensuring that the injection depth is within the appropriate range, and improving the uniformity and safety of the injection effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118936308A_ABST
    Figure CN118936308A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent identification method for live pig injection, relates to the technical field of intelligent live pig injection, and solves the problems that the thickness of pigskin in a live pig injection area cannot be accurately detected in the prior art, and the injection technology mostly adopts fixed pressure injection or injection by manually adjusting injection pressure for injection; injection parameters cannot be dynamically adjusted according to the thickness of pigskin in an injection area, so that vaccine injection is not uniform or the injection depth is insufficient, and the immune effect is affected. The method comprises the following steps: acquiring three-dimensional point cloud data of live pigs by using a laser beam irradiation technology; segmenting the overall point cloud data of the live pig into different areas through a point cloud segmentation algorithm to obtain a plurality of areas to be identified, if the areas to be identified are injection areas, calculating the pigskin thickness of the injection areas, and adjusting the injection parameters of an injector to inject the injection areas; the injection parameters of the injector are dynamically adjusted, it is ensured that the injection depth is within a suitable range, and the good injection effect is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of intelligent injection of live pigs, and specifically relates to a method for intelligent identification of injection of live pigs. Background Art

[0002] In the intelligent pig farm, each pig will be identified and recorded through specific technical means (such as RFID technology), which provides basic data support for intelligent pig identification injection. For pig immunization, injection is usually performed with fixed pressure injection or manually adjusted injection pressure; among them, fixed pressure injection is not effective for pigs at different stages, and the other is to manually adjust the pressure for injection, which needs to be adjusted according to the operator's experience. In the early stage of injection, there is a situation where the injection effect is not ideal due to inadequate adjustment, and it is impossible to achieve one pressure for each pig, which affects the injection effect.

[0003] In addition, the immune effect of existing needle-free injections is affected by many factors, including injection distance, injection angle, injection speed, and the location and skin thickness of the pig. Among them, the detection of pig skin thickness is particularly important. The existing methods for detecting pig skin thickness include probe method and ultrasonic method. Among them, the probe method is to use a metal ruler to directly insert into the measurement location of the pig to measure the vertical distance from the skin to the muscle or fat layer, thereby obtaining skin thickness data. This method is harmful to live pigs and is easy to bring bacteria and viruses into the live pigs. It has high requirements for operators and is inconvenient to operate. It is easy to cause stress reactions in live pigs, and operators have safety risks. The measurement results are greatly affected by the operator's experience and technique, and there may be certain errors. The ultrasonic method uses the propagation speed and reflection characteristics of ultrasonic waves between different tissue layers (such as skin, fat, muscle, etc.) in the pig's body to measure skin thickness. When ultrasonic waves encounter different tissue layers, reflection and scattering will occur. By receiving and analyzing these reflected signals, the skin thickness data can be calculated. This method requires professional equipment and technical support, which is relatively costly. In addition, a coupling agent must be applied before measurement, and the measurement process must be performed strictly in accordance with the operating procedures. There are many measurement error factors. Therefore, the present invention provides a method for intelligent identification of pig injections. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a method for intelligent identification of pig injections, which is used to solve the technical problem that the prior art cannot accurately detect the thickness of the pig skin in the injection area of ​​the pig, and the injection technology is mostly fixed pressure injection or manually adjusted injection pressure for injection, and the injection parameters cannot be dynamically adjusted according to the thickness of the pig skin in the injection area, resulting in uneven vaccine injection or insufficient injection depth, thereby affecting the immune effect.

[0005] To achieve the above object, the first aspect of the present invention provides a method for intelligent identification of pig injections, comprising the following steps:

[0006] Step 1: Use a laser beam to illuminate the surface of the pig, and calculate the three-dimensional coordinate information of the pig surface by measuring the round-trip time and angle change of the laser beam to form three-dimensional point cloud data;

[0007] Step 2: Segment the overall point cloud data of the pig into different areas through the point cloud segmentation algorithm to obtain several areas to be identified, and input them into the pig identification model to match the areas to be identified with the corresponding parts of the pig; determine whether the area to be identified is the injection area; if yes, proceed to step 3; if not, do nothing;

[0008] Step three: emit laser pulses to the injection area, receive reflected light, and calculate the pig skin thickness in the injection area based on the round-trip time of the laser beam; obtain the injection parameters of the syringe based on the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and adjust the injection parameters of the syringe to inject the injection area; wherein the basic information of the pig includes the breed, age and gender of the pig.

[0009] Preferably, the method of calculating the three-dimensional coordinate information of the surface of the pig by measuring the round-trip time and angle change of the laser beam includes:

[0010] By rotating the laser head, measurements are taken at different angles, and the angles of the laser beam at different positions are recorded, and the one-way propagation distance of the laser beam is calculated; wherein the angle includes the vertical angle and the horizontal angle; the one-way propagation distance S = c × t1 / 2; c is the propagation speed in vacuum, and t1 is the round-trip time;

[0011] By formula The spatial coordinates (x, y, z) of the measurement points are calculated, and the spatial coordinates of all measurement points are combined to form three-dimensional point cloud data.

[0012] First, the present invention realizes all-round and multi-angle measurement by rotating the laser head, which greatly improves the comprehensiveness and accuracy of data acquisition. Compared with the traditional single-point measurement method, this method can quickly capture a large number of details on the surface of the object, providing a rich data source for subsequent three-dimensional reconstruction or spatial analysis; secondly, the one-way propagation distance of the laser beam is calculated by the reflection principle, which effectively eliminates the errors that may be introduced in the round-trip time measurement, ensuring the accuracy of the distance data. High-precision measurement is the basis for building an accurate three-dimensional model; finally, the angle and distance information are converted into spatial coordinates through the spherical coordinate conversion formula, realizing the conversion from original measurement data to three-dimensional point cloud data.

[0013] Preferably, the pig identification model comprises:

[0014] A series of labeled three-dimensional point cloud data are collected from historical data. Each three-dimensional point cloud data contains a complete pig model, and each part of the pig model is labeled separately, and input into the CNN neural network model for training to obtain a pig recognition model. Among them, the input data of the CNN neural network model is the three-dimensional point cloud data, and the output data is the label corresponding to the pig part area, and the label is set based on natural numbers.

[0015] The present invention ensures the richness and diversity of training data by collecting labeled three-dimensional point cloud data from historical data. The data set based on real scenes can more accurately reflect the complex morphology and changeable postures of pigs, providing a solid foundation for model learning; fine labeling of various parts of the pig model not only improves the recognition accuracy of the model, but also enables the model to understand the structural characteristics of the pig in more detail; and the CNN neural network model is used as a training framework, making full use of its powerful ability in processing images and point cloud data. CNN can automatically extract deep features in the data and build an effective recognition model by learning the complex relationship between these features, thereby realizing accurate recognition of various parts of the pig.

[0016] Preferably, the method of calculating the thickness of the pigskin in the injection area according to the round trip time and angle change of the laser beam comprises:

[0017] The thickness d of the pig skin in the injection area is calculated by the formula d = (c×t2) / (2×n×cosγ); where c is the speed of light in a vacuum, t2 is the round-trip time of the laser pulse, n is the refractive index of the pig skin in the injection area to the laser, and γ is the incident angle.

[0018] The present invention irradiates the injection area of ​​the pig with a laser beam and uses the propagation characteristics of the laser pulse in the pig skin to indirectly measure the thickness of the pig skin. It has the characteristics of non-contact and non-destructiveness, which is particularly important for scenarios where the health status of the pigs needs to be frequently monitored or precision medical operations need to be performed, and can avoid causing additional pain or stress reactions to the pigs. Secondly, considering the influence of the incident angle γ on the actual propagation distance of light in the pig skin, the method accurately calculates the propagation distance in the vertical direction by introducing a trigonometric function relationship, and accurately evaluates the pig skin thickness in the injection area by using the ratio of the propagation distance in the vertical direction to the refractive index, laying a foundation for adjusting the injection parameters of the syringe.

[0019] Preferably, the step of obtaining the injection parameters of the syringe according to the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area includes:

[0020] An empirical relationship model is established through experiments and statistical analysis and consulting with veterinarians or breeding experts. The relationship model outputs the parameters of the syringe based on the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area; the input data of the model includes the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and the output data includes the needle size, injection depth and injection pressure.

[0021] The present invention establishes the relationship between the injection area and the injection parameters of the syringe through a large amount of experimental data, ensures the scientificity and rigor of the model establishment, and introduces the opinions of veterinary or breeding experts, so that the model is not only based on data, but also integrates industry experience and professional knowledge. The participation of experts can ensure the effectiveness and practicality of the model in practical applications, avoiding the deviations or limitations that may be caused by relying solely on data; by accurately adjusting the parameters of the syringe, the success rate and safety of the injection are improved.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. The present invention irradiates the surface of a pig with a laser beam to construct three-dimensional point cloud data of the pig, labels each part of the three-dimensional point cloud data, and inputs it into a CNN neural network model for training to obtain a pig recognition model. The CNN model can learn the characteristics of different areas of the pig to achieve more accurate recognition; in addition, by emitting laser pulses and receiving their reflected light, the thickness of the pig skin is calculated according to the round-trip time and angle change of the laser beam, thereby achieving high-precision measurement of the thickness of the live pig skin. This method does not require additional consumables and is easy to operate;

[0024] 2. The present invention utilizes a laser beam to irradiate the surface of a live pig, calculates the three-dimensional coordinate information of the live pig surface by measuring the round-trip time and angle change of the laser beam, and forms three-dimensional point cloud data; the overall point cloud data of the live pig is segmented into different areas by a point cloud segmentation algorithm to obtain a number of areas to be identified, and inputs them into a live pig identification model to match the areas to be identified with the corresponding parts of the live pig; if the area to be identified is an injection area, the pig skin thickness of the injection area is calculated; the injection parameters of the syringe are dynamically adjusted according to the basic information of the pig, the injection area, and the pig skin thickness of the injection area to ensure that the injection depth is within a suitable range to achieve a good injection effect; the present invention realizes accurate identification of various parts of live pigs and precise measurement of pig skin thickness, providing strong support for the fields of intelligent pig farming, veterinary diagnosis, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only 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.

[0026] Figure 1 It is a schematic diagram of the process of the present invention;

[0027] Figure 2 This is a schematic diagram of the process of the pig part area identification method of the present invention;

[0028] Figure 3 The figure is a schematic flow chart of the method for analyzing pigskin thickness of the present invention. DETAILED DESCRIPTION

[0029] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only 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.

[0030] See also Figure 1 The first aspect of the present invention provides a method for intelligent identification of pig injections, comprising the following steps:

[0031] Step 1: Use a laser beam to illuminate the surface of the pig, and calculate the three-dimensional coordinate information of the pig surface by measuring the round-trip time and angle change of the laser beam to form three-dimensional point cloud data;

[0032] Specifically, by rotating the laser head, measurements are performed at different angles, and the angles of the laser beam at different positions are recorded, and the one-way propagation distance of the laser beam is calculated; wherein the angle includes the vertical angle and the horizontal angle; the one-way propagation distance S = c × t1 / 2; c is the propagation speed in vacuum, and t1 is the round-trip time;

[0033] By formula The spatial coordinates (x, y, z) of the measurement points are calculated, and the spatial coordinates of all measurement points are combined to form three-dimensional point cloud data.

[0034] For example: The laser head of a laser scanner rotates in horizontal and vertical directions to cover the required measurement area. The laser head emits laser pulses at multiple preset angles to capture different parts of the target surface and record the horizontal angle of the laser beam relative to the instrument reference plane. And the vertical angle θ. Among them, the horizontal / vertical angle is usually measured using a high-precision rotary encoder or angle sensor. By measuring the time from the laser pulse being emitted to returning to the receiver, the round-trip time t1 of the laser pulse is calculated; since the round-trip time is the two-way time, the one-way time is t1 / 2.

[0035] The spatial coordinates of the measuring points are calculated based on the distances and horizontal / vertical angles obtained from the measuring points, and the spatial coordinates of all the measuring points are combined to form a three-dimensional point cloud data set, that is, a three-dimensional model of the pig is obtained.

[0036] Step 2: Segment the overall point cloud data of the pig into different areas through the point cloud segmentation algorithm to obtain several areas to be identified, and input them into the pig identification model to match the areas to be identified with the corresponding parts of the pig; determine whether the area to be identified is the injection area; if yes, proceed to step 3; if not, do nothing;

[0037] See also Figure 2 Specifically, a series of labeled three-dimensional point cloud data are collected from historical data, each of which contains a complete pig model, and each part of the pig model is labeled separately, and input into the CNN neural network model for training to obtain a pig recognition model; wherein, the input data of the CNN neural network model is the three-dimensional point cloud data, and the output data is the label corresponding to the pig part area, and the label is set based on natural numbers.

[0038] For example: extract a number of 3D point cloud data from historical data, and manually or semi-automatically annotate different parts of each 3D point cloud data, such as ears, tails, legs, etc. Each part is assigned a specific digital label, such as ears are marked as 1, tails are marked as 2, etc. The labeled data set is divided into a training set and a test set, and the training set is used to train the CNN neural network. The verification machine is used to test the effect of the training, improve the accuracy and robustness of the model, and obtain a pig recognition model.

[0039] Step 3: Emit laser pulses to the injection area, receive reflected light, and calculate the pig skin thickness in the injection area based on the round-trip time of the laser beam;

[0040] See also Figure 3 Specifically, the thickness d of the pig skin in the injection area is calculated by the formula d = (c×t2) / (2×n×cosγ); wherein c is the speed of light in a vacuum, t2 is the round-trip time of the laser pulse, n is the refractive index of the pig skin in the injection area to the laser, and γ is the incident angle.

[0041] It should be noted that the refractive index of the present invention is a known constant value and does not change with the change of laser wavelength. The incident angle γ is the angle between the laser beam and the normal line, and the normal line is the laser beam perpendicular to the surface of the injection area of ​​the pig.

[0042] The measurement process of the incident angle is as follows:

[0043] Determine a reference plane, usually a horizontal plane or a tangent plane of the pig surface, determine a normal direction perpendicular to the reference plane, fix the laser emitter in a stable position, ensure that the direction of the laser beam is stable and adjustable, so that the laser beam it emits can irradiate the pig surface, and use an angle measurement tool, such as a protractor or an optical angle measurement device, to measure the angle between the laser beam and the normal, that is, to obtain the size of the incident angle.

[0044] The present invention irradiates the injection area of ​​the live pig with a laser beam, and calculates the propagation distance of the laser pulse in the pig skin in one round trip. The incident angle γ affects the actual propagation distance of the light in the pig skin. When the light is incident at a certain angle, its actual propagation distance in the pig skin will be longer than when it is incident vertically. In order to calculate the propagation distance in the vertical direction, the incident angle γ needs to be considered, that is, the ratio of the laser one-way propagation distance to cosγ is used to obtain the propagation distance in the vertical direction. The ratio of the propagation distance in the vertical direction to the refractive index is used to obtain the pig skin thickness in the injection area.

[0045] Furthermore, the injection parameters of the syringe are obtained according to the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and the injection parameters of the syringe are adjusted to inject the injection area; wherein the basic information of the pig includes the breed, age and gender of the pig.

[0046] Specifically, an empirical relationship model is established through experiments and statistical analysis and consulting with veterinarians or breeding experts. The relationship model outputs the parameters of the syringe based on the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area; wherein the input data of the model includes the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and the output data includes the needle size, injection depth and injection pressure.

[0047] For example: 1. Data collection: The basic information of the pig includes age, weight, gender, breed, etc., injection areas such as behind the ear, shoulder, buttocks, etc., and the thickness of the pig skin in the corresponding injection area. 2. Experimental design: Select representative pig samples to cover various basic information combinations; Syringe parameters: Try different needle sizes, injection depths and injection pressures in the experiment based on the pig information. 3. Experimental implementation: Perform injection operations according to the set parameters and record relevant data; 4. Observation records: Observe the reactions after injection, such as whether the injection is successful, whether there are adverse reactions, etc., and record them, and keep records of successful injections. 5. Feedback collection: Collect the opinions of veterinarians and breeding experts to make reasonable adjustments to the parameters.

[0048] A relational model is established based on the retained data to output the optimal needle size, injection depth and injection pressure according to the input basic information of the pig, injection area and pig skin thickness.

[0049] The relationship model may be a linear regression model, a decision tree, a random forest, or other types of machine learning models.

[0050] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0051] The working principle of the present invention is as follows: a laser beam is used to irradiate the surface of a live pig, and the three-dimensional coordinate information of the live pig surface is calculated by measuring the round-trip time and angle change of the laser beam to form three-dimensional point cloud data; the overall point cloud data of the live pig is divided into different areas by a point cloud segmentation algorithm to obtain a plurality of areas to be identified, and the areas are input into a live pig identification model to match the areas to be identified with the corresponding parts of the live pig; if the part is an injection area, a laser pulse is emitted to the injection area, and reflected light is received, and the pig skin thickness of the injection area is calculated according to the round-trip time of the laser beam; the basic information of the pig, the injection area, and the pig skin thickness of the injection area are input into a relational model, and the relational model outputs the injection parameters of the syringe, and the injection parameters of the syringe are adjusted to inject the injection area; wherein the basic information of the pig includes the breed, age and gender of the pig.

[0052] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for intelligent identification of pig injections, It is characterized in that The following steps are involved: Step 1: Use a laser beam to illuminate the surface of the pig, and calculate the three-dimensional coordinate information of the pig surface by measuring the round-trip time and angle change of the laser beam to form three-dimensional point cloud data; Step 2: Segment the overall point cloud data of the pig into different areas through the point cloud segmentation algorithm to obtain several areas to be identified, and input them into the pig identification model to match the areas to be identified with the corresponding parts of the pig; determine whether the area to be identified is the injection area; if yes, proceed to step 3; if not, do nothing; Step three: emit laser pulses to the injection area, receive reflected light, and calculate the pig skin thickness in the injection area based on the round-trip time of the laser beam; obtain the injection parameters of the syringe based on the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and adjust the injection parameters of the syringe to inject the injection area; wherein the basic information of the pig includes the breed, age and gender of the pig.

2. A method for intelligent identification of pig injections according to claim 1, It is characterized in that The method of calculating the three-dimensional coordinate information of the surface of the pig by measuring the round trip time and angle change of the laser beam includes: By rotating the laser head, measurements are taken at different angles, and the angles of the laser beam at different positions are recorded, and the one-way propagation distance of the laser beam is calculated; wherein the angle includes the vertical angle and the horizontal angle; the one-way propagation distance S = c × t1 / 2; c is the propagation speed in vacuum, and t1 is the round-trip time; By formula The spatial coordinates (x, y, z) of the measurement points are calculated, and the spatial coordinates of all measurement points are combined to form three-dimensional point cloud data.

3. A method for intelligent identification of pig injection according to claim 2, It is characterized in that The pig identification model comprises: A series of labeled three-dimensional point cloud data are collected from historical data. Each three-dimensional point cloud data contains a complete pig model, and each part of the pig model is labeled separately, and input into the CNN neural network model for training to obtain a pig recognition model. Among them, the input data of the CNN neural network model is the three-dimensional point cloud data, and the output data is the label corresponding to the pig part area, and the label is set based on natural numbers.

4. A method for intelligent identification of pig injections according to claim 1, It is characterized in that The method of calculating the thickness of the pig skin in the injection area according to the round trip time and angle change of the laser beam includes: The thickness d of the pig skin in the injection area is calculated by the formula d = (c×t2) / (2×n×cosγ); where c is the speed of light in a vacuum, t2 is the round-trip time of the laser pulse, n is the refractive index of the pig skin in the injection area to the laser, and γ is the incident angle.

5. A method for intelligent identification of pig injections according to claim 4, It is characterized in that The method of obtaining the injection parameters of the syringe according to the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area includes: An empirical relationship model is established through experiments and statistical analysis and consulting with veterinarians or breeding experts. The relationship model outputs the parameters of the syringe based on the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area; the input data of the model includes the basic information of the pig, the injection area, and the thickness of the pig skin in the injection area, and the output data includes the needle size, injection depth and injection pressure.