| Execution Methods |
- This project develops hatching systems and technologies with visual recognition and intelligent diversion for the hatching of fertilized chicken eggs. This system includes a system link of three technologies: (1) thermographic images for chicken fertilized egg biological data collection, (2) visual recognition deep learning platform and (3) artificial intelligence classification equipment on fertilized eggs.
- The purpose of this project includes (1) To collect the biological temperature perception data of different stages on fertilized eggs in the incubator by thermometer. To visually identify and distinguish the hatching egg process, and to establish the incubation conditions for different grades of fertilized eggs. (2) To establish automatic shunt equipment for classified fertilized eggs, reduce the loss caused by manual operation and shorten operation time exhaustion. (3) Reducing the stress of hatching chicks in the incubator and transportation.
- Using the above technical links to complete: (1) intelligent all-in and all-out fertilized eggs, (2) intelligent automatic shunting and transmission of fertilized eggs, (3) connecting intelligent incubation, automatic visual recognition deep learning and transportation, and (4) advanced on-farm hatching model.
|
| Performance Evaluation |
Specific results: - Using infrared thermometer for various temperature images selection on biological characteristics from hatching fertilized chicken eggs. The infrared thermographic data were real-time analyzed by designed automatic detection equipment linked with fertilized eggs incubator at the same time. The proportion of self-made works at this stage is 100%.
- Using Mask R-CNN and convolutional neural network fertilized egg temperature image object detection model to complete the accurate classification of whole fertilized egg incubation period.
- Transfer the coordinate information of the target fertilized egg identified by the image recognition system to the vacuum suction cup equipment for sorting and shunting.
Activity effects: - Using intelligent visual recognition technology to accurately detect and improve the hatching rate of fertilized eggs of breeders.
- Based on the autonomous heat production characteristics of the new hatching chicks, using thermometer to detect of fertilized eggs and link with intelligent visual recognition technique to accurately standardize the separation of eggs contained hatching chicks, so as to achieve “intelligent AI for all-in and all-out fertilized eggs”.
|
| Conclusion & Suggestion |
The developed intelligent AI visual recognition and intelligent hatching fertilized eggs shunting system have been linked to incubator and feeding model of the chick farm. These uniformly standardized hatching eggs can be hatched and bred in a farm with constant temperature and humidity control. These new hatched chicks can directly eat starter diet and drink water to reduce the related stress, shorten the breeding time of the broilers, and construct a precision agricultural production system. At present, we has cooperated with the hatchery chicken industry, the chicken farm breeding equipment industry and the automatic diversion equipment industry to link the relevant devices to the stage of practical application. In addition, our 2022 new project “Conduct a novel intelligent agriculture AI vision recognition of duck fertilized eggs” has been approved by Council of Agriculture Executive Yuan. This new project is cooperated with Yilan branch, Livestock Research Institute, COA to work on intelligent classification of duck fertilized eggs. It is an application of the technology derived from this project. It is suitable for the promotion of industry-government-academic cooperation which owned characteristics of local poultry intelligent breeding. |
| Appendix |
Appendix 1 |
