CN-122023659-A - Livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration

CN122023659ACN 122023659 ACN122023659 ACN 122023659ACN-122023659-A

Abstract

The invention provides a digital twin construction and dynamic calibration method of livestock and poultry individuals based on body ruler and weight double calibration, which relates to the field of intelligent cultivation and agricultural information engineering, the invention intelligently screens effective dynamic weight values and eliminates abnormal data based on weight change rate and three-dimensional gesture analysis by synchronously collecting dynamic weighing data and three-dimensional point cloud data of the livestock and poultry individuals, calibrates the effective dynamic weight and static reference weight in a short time window to generate weight correction factors, weighting and fusing the multiple historical correction factors according to the acquisition time and the stability to obtain a fusion weight correction parameter, finally driving the three-dimensional digital twin model to update through the biometric consistency constraint based on the parameter and the latest body size data, the method and the device effectively improve the precision, the anti-interference capability and the long-term reliability of the digital twin model, and are suitable for accurate cultivation management of livestock and poultry individuals.

Inventors

LEI KAIDONG
LIU XIANGLONG
LI YUECHAO
ZHANG SHIQI
GUO ZHENZHEN
LI KUNPENG

Assignees

山西农业大学

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (8)

1. A livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration is characterized by comprising the following steps: synchronously acquiring dynamic weighing data and three-dimensional point cloud data of an individual target livestock and poultry, and constructing an initial parameterized three-dimensional digital twin model based on the acquired data; Step two, processing the collected dynamic weighing data, and screening out effective dynamic weight values based on a preset stability judgment rule; calibrating the effective dynamic weight value and the static reference weight value of the same target livestock and poultry individual acquired in a preset time window to generate a weight correction factor; Step four, based on a plurality of weight correction factors generated at different times for the same target livestock and poultry individuals, carrying out weighted fusion to generate fusion weight correction parameters; step five, updating the parameterized three-dimensional digital twin model according to the fusion weight correction parameters and body ruler parameters extracted from the latest three-dimensional point cloud data; And step six, repeatedly executing the step two to the step five on the basis of the updated parameterized three-dimensional digital twin model, and realizing iterative calibration and state tracking of the model.
2. The method for constructing and dynamically calibrating individual digital twin of livestock and poultry based on body ruler and body weight double calibration according to claim 1 is characterized in that in the first step, the parameterized three-dimensional digital twin model is generated in the specific mode that: Continuously acquiring real-time weight sequence data of a target livestock and poultry individual through an external dynamic weighing device, acquiring three-dimensional point cloud data of the target livestock and poultry individual through an external three-dimensional vision device, and generating an initial parameterized three-dimensional digital twin model based on an initial average value of the real-time weight sequence data and body size parameters extracted from the three-dimensional point cloud data.
3. The method for constructing and dynamically calibrating digital twin of livestock and poultry based on body size and weight double calibration according to claim 2 is characterized in that in the second step, the specific mode of screening effective dynamic weight values is as follows: Calculating the change rate of the real-time weight sequence data, judging to enter a weighing stable interval when the change rate is continuously lower than a preset change rate threshold value and the state duration time lower than the change rate threshold value reaches a preset stable duration time threshold value, recording the duration time of the stable interval, and marking the average value of the weight data in the interval as an effective dynamic weight value.
4. The method for digital twin construction and dynamic calibration of livestock and poultry individuals based on body ruler and body weight double calibration according to claim 3, wherein in the second step, the method further comprises the following steps: Based on the three-dimensional point cloud data, calculating the space displacement or attitude angle of key points of the target livestock and poultry individuals, and when the space displacement or attitude angle exceeds a preset attitude tolerance threshold, judging that the target livestock and poultry individuals are in an unstable motion attitude, discarding data corresponding to the acquisition window, and not marking the data as an effective dynamic weight value.
5. The method for constructing and dynamically calibrating the digital twin of the livestock and poultry based on the body size and the weight double calibration of the livestock and poultry according to claim 1 is characterized in that in the third step, the specific mode for generating the weight correction factors is as follows: And acquiring a static reference weight value of the target livestock and poultry individual within a preset time window after the effective dynamic weight value is acquired, calculating a difference value of the static reference weight value of the effective dynamic weight value, and taking the difference value as a weight correction factor.
6. The method for constructing and dynamically calibrating the digital twin of the livestock and poultry based on the body size and weight double calibration of the invention as claimed in claim 1, wherein in the step four, the specific method of weighting fusion is as follows: and distributing a fusion weight for each weight correction factor, wherein the fusion weight is determined based on the data acquisition time corresponding to the weight correction factor and the duration time of the weighing stable interval corresponding to the factor when the factor is generated, and then carrying out weighted average calculation on all the weight correction factors according to the fusion weights corresponding to the weight correction factors to obtain the fusion weight correction parameters.
7. The method for constructing and dynamically calibrating the digital twin of the livestock and poultry based on the body size and the weight double calibration of the livestock and poultry according to claim 1 is characterized in that in the fifth step, the updating concrete mode is as follows: And acquiring a body scale-weight biological statistical model of the species to which the target livestock and poultry individuals belong, inputting the latest extracted body scale parameters into the biological statistical model to obtain a weight predicted value, calculating the deviation between the corrected weight corresponding to the fusion weight correction parameter and the weight predicted value, and carrying out three-dimensional geometric deformation on the parameterized three-dimensional digital twin model according to the deviation when the deviation exceeds a preset tolerance range, and simultaneously adjusting the quality attribute of the parameterized three-dimensional digital twin model according to the fusion weight correction parameter.
8. The method for constructing and dynamically calibrating individual digital twins of livestock and poultry based on body ruler and body weight double calibration according to claim 1, wherein in the step six, the updated parameterized three-dimensional digital twins model is used as the model version based on the next iteration execution step two to step five.

Description

Livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration Technical Field The invention relates to the technical field of intelligent breeding and agricultural information engineering, in particular to a livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration. Background In modern livestock and poultry cultivation, the realization of accurate management on individual levels is a key for improving cultivation efficiency and guaranteeing animal welfare. The body weight and body size (e.g., body length, height, chest circumference) of an individual are core physiological indicators reflecting their growth performance, nutritional status, and health level. The traditional cultivation management relies on manual periodic spot check to carry out weighing and body size measurement, and the method has the advantages of high labor intensity, low measurement frequency, easy stress on animals and difficulty in obtaining a continuous individual growth curve. In addition, the manual measurement error is larger, and the data consistency is poor. With the development of the internet of things and machine vision technology, attempts to automatically measure by dynamic weighing orbit or visual estimation and other methods are presented. However, the prior art scheme is often limited in that the dynamic weighing is simply relied on and is easily influenced by the movement state (such as walking speed change, jump and pause) of animals when the animals pass, so that the data fluctuation is large and the reliability is low, the body ruler and the weight are estimated only through two-dimensional or three-dimensional vision, the accuracy is obviously influenced by illumination, shielding and animal posture, the direct physical calibration with the real weight is lacking, and the long-term prediction deviation of the model can be gradually accumulated. The method is difficult to construct a digital twin model capable of continuously and accurately reflecting the state change of a real individual, and cannot meet the high requirement of accurate feeding and health early warning of the individual on long-term consistency of data, so that the invention provides a livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration to solve the problems in the prior art. Disclosure of Invention Aiming at the problems, the invention aims to provide a livestock and poultry individual digital twin construction and dynamic calibration method based on body ruler and body weight double calibration, and aims to introduce an effective data screening mechanism based on motion state perception, a calibration mechanism from dynamic data to static reference and a weighted fusion mechanism of a cross-period historical calibration result through synchronous acquisition and fusion processing of dynamic weighing and three-dimensional vision, so as to finally realize automatic construction, high-precision updating and long-term state consistency maintenance of a livestock and poultry individual parameterized three-dimensional digital twin model and provide a reliable data basis for accurate individual cultivation. The invention aims at realizing the aim by adopting the following technical scheme that the method for constructing and dynamically calibrating the digital twin of the individual livestock and poultry based on the body ruler and the body weight double calibration comprises the following steps: synchronously acquiring dynamic weighing data and three-dimensional point cloud data of an individual target livestock and poultry, and constructing an initial parameterized three-dimensional digital twin model based on the acquired data; Step two, processing the collected dynamic weighing data, and screening out effective dynamic weight values based on a preset stability judgment rule; calibrating the effective dynamic weight value and the static reference weight value of the same target livestock and poultry individual acquired in a preset time window to generate a weight correction factor; Step four, based on a plurality of weight correction factors generated at different times for the same target livestock and poultry individuals, carrying out weighted fusion to generate fusion weight correction parameters; step five, updating the parameterized three-dimensional digital twin model according to the fusion weight correction parameters and body ruler parameters extracted from the latest three-dimensional point cloud data; And step six, repeatedly executing the step two to the step five on the basis of the updated parameterized three-dimensional digital twin model, and realizing iterative calibration and state tracking of the model. The method is further improved in that in the first step, the parameter