Animal activity monitoring system and activity state recognition method based on neural network

Abstract

The invention discloses an animal activity monitoring system and an activity state recognition method based on a neural network. The animal activity monitoring system comprises an acquisition device and an intelligent terminal application, and the acquisition device acquires contact pressure, activity range and acceleration information of monitored animals and judges whether effective contact occurs or not according to the magnitude of the contact pressure; the intelligent terminal application receives effective animal data of the acquisition device and recognizes the effective animal data into the animal activity state. By means of the recognition method, based on a neural network model, a sample set can be trained and learned, and based on the trained model, the activity state of the animals can be recognized. By means of the system, animal activities can be automatically detected and remotely monitored in real time, manual detection or recording is not needed, and a lot of manpowercan be saved; by means of the method, the difference like the type, the age, the individual and the motion state of the animals can be eliminated, all monitored objects are truly customized, and the purpose of precisely and effectively detecting the activity range is achieved.

Description

technical field [0001] The invention relates to the field of animal health monitoring, in particular to an animal activity monitoring system and a neural network-based activity state identification method. Background technique [0002] Activity is an important health indicator of animals. Activity monitoring is one of the important ways to evaluate the health status of animals. The timeliness, accuracy and reliability of activity monitoring directly affect the prevention, diagnosis, treatment and postoperative recovery of diseases. [0003] It is very inconvenient to track changes in the range of activities of animals in the traditional way, and the current measurement of the range of activities of animals is mainly done manually. The owner needs to manually measure the range of activities of animals every day, record the distance of activities, and draw the curve of the range of activities. This is a big problem. workload. Moreover, some animals, such as grazing cattle, ca...

AI Technical Summary

Problems solved by technology

[0003] Traditional animal range of motion is very inconvenient to track changes in animal range of motion, and the current measurement of animal range of motion is mainly done manually. The owner needs to manually measure the range of animals every day, record the distance of motion, and draw the curve of range of motion, which is a big challenge. workload

Moreover, some animals, such as grazing cattle, cannot cooperate with the range of activity detection. For such animals, traditional activity detection methods cannot implement high-frequency detection

[0004] In addition, different kinds of animals, different age groups of the same kind of animals, different individuals of the same kind of animals at the same age, and different movement states of the same individual animals often have different ranges of movement. Temporarily increase the range of activities; another example: the range of activities of adult dogs is relatively large, while the normal range of activities of puppies is small. Therefore, the monitoring of animal activities should take into account the differences in animal species, age, individual, and exercise status. It is difficult to fully achieve the traditional method

Images