Font size:
Small
Medium
Large

The intelligence development of livestock industry

Accord to the paper of 2020 Agricultural Statistics Yearbook, there is 16.67 billion NT dollars of livestock production and reaching 33.5% of agriculture production in Taiwan. Production of livestock to sort was 1st hogs, 2nd broilers, 3rd layers, 4th cow milk and 5th mule & muscovery ducks and these productions were 71.4, 46.4, 21.4, 11.4 and 7.5 billion NT dollars, respectively. Accounting the year production were 10.41 and 19.75 million were hogs farm and cow milk farm. But also these are labor shortage on the farm. The younger farmers of livestock industry these agree and want to use 10% income of farm to use intelligent machines on farm and to substitute for labor.

The last few year results of study on intelligent livestock were showed those of The Intelligent cooling and ventilation prototype system of nursery pig house, Optimization the care system of sow, Automatic buffalo breeding management demonstration system, Cow calving voiceprint early warning system, Three-dimensional image capture system for the body measurement in pig and Sow hat recognition system. These projects studied on the Intelligent cooling and ventilation prototype system of nursery pig house: There were used the LoRaWAN wireless transmission network structure to connect with the office computer. The fan can be turned on and off according to by with the temperature, humidity. The relevant data is stored in the online database. The results showed that of the system can significantly reduce the environmental temperature and humidity index (THI), improve the heat stress of pigs, and raise feed efficiency. However, it is necessary to continue to collect basic data, and possible to install sprinkler facilities and harmful gas sensors for monitoring barn environment precisely for livestock house and precision feed, sow heat recognition system and artificial intelligent for feeding and management to solve and reduce the cost of labor on the farm.

Title

Introduction

Three-dimensional image capture system for the body measurement in pig

Taiwan Livestock Research Institute (TLRI) cooperated with the Industrial Technology Research Institute (ITRI) to develop the 3D image capture system for pig body measurement. This system provides the sense of animal size, including a frame-type container module, multiple three-dimensional imaging modules, and using a computer system. The three-dimensional imaging modules capture multiple point clouds of a target animal from different angles. The computer system obtains a depth map viewed from the side according to point clouds, compensates for railing parts, and calculates the waist or chest depth of the target animal. This system successfully obtained the information on a pig conformation within 15 seconds and continued to collect the shape data from pigs auctioned in the Central Performance Test Station. In the future, this system will effectively develop outstanding boars and expand the supply of high-quality pork in meat production.

Sow hat recognition system

The intelligent sow estrus monitoring system was established by the collection of basic data on the sow's heat behavior and the degree of swelling of the vulva, and then the image recognition system was used to measure the height and width of the sow's vulva. According to the heat behavior of the sows on the day of estrus was more frequent and the time is longer, it was judged when the sows entered the stable mating period and had the behavior of looking for boars. This system could increase the success rate of sow mating, reduce feeding costs, accurately know the optimal mating time and shorten the days open period of sows, and thereby improving the production efficiency of pig farms.

Intelligent monitoring system for sow farrowing house

This research department is jointly developed with the Microsystem Center of ITRI to assist in the development. The system includes:1. Temperature and humidity sensors (environmental perception),2. Radio monitoring equipment (sound/emotion recognition), 3. Video monitoring equipment (motion behavior analysis) ) and 4. Thermal imaging monitoring system (physiological/behavioral recognition), it can be used in sow farrowing houses, through the collection and recognition of images and voiceprints, all-weather assisted monitoring, observation, and detection of sows close to the clinic before delivery for signs, immediately notify the management staff to provide appropriate assistance, as well as prevent piglets from being crushed after delivery and lactation, so as to improve the breeding rate of piglets.

Intelligent cooling and ventilation prototype system of nursery pig house

We build with fan, temperature and humidity sensors in the nursery pig house. There were used the LoRaWAN wireless transmission network structure to connect with the office computer. The fan can be turned on and off according to by with the temperature, humidity. The relevant data is stored in the online database. The results showed that of the system can significantly reduce the environmental temperature and humidity index (THI), improve the heat stress of pigs, and raise feed efficiency. In the future, it is necessary to continue to collect basic data, and possible to install sprinkler facilities and harmful gas sensors for monitoring barn environment precisely.

Automatic buffalo breeding management demonstration system

The data collection of body weight, body shaped and the number of buffalo can be easily obtained by connecting RFID weight scale, RFID multiple scanners, PDA with special software.

Cow calving voiceprint early warning system

This project collected the voiceprint data of cows' moo and found that the pre-parturition calls of cows are lower in audio frequency and longer than that of the normal cow's calls. In order to enhance the accuracy rate of distinguishing the cow moo voiceprint features, we used algorithm technology to strengthen and validate the features. Now we could distinguish the difference between the birthing cow moo or normal sounds and the accuracy rate reached 90%.
TOP