Author：Chih-Kai Chiu,Industrial Technology Research Institute
Chi-Yuan Lin,Fisheries Research Institute, COA
Yen-Ting Lin,Creative Way Global Co., Ltd.

According to the estimates of Food and Agriculture Organization, the global output value of ornamental fish is over 15 billion. Southeast Asia has an excellent breeding environment and geographical location, with low production and transportation costs to seize export opportunities. However, Taiwan's production and marketing information integration is insufficient, and it is essential to establish digital information flow and intelligent breeding method.

The demonstration company of Creative Way Global Co., Ltd. (CWG) has been committed to exporting Taiwan's ornamental fish, such as freshwater fish, saltwater fish, water plant, and coral to markets around the world. CWG is also Taiwan’s leading trader of ornamental fish. With the assistance of the Smart Agriculture Project, CWG applies the technology of “artificial intelligence recognition and control system for ornamental fish length” developed by Fisheries Research Institute, COA and Industrial Technology Research Institute. In addition, CWG combines the visual counting technology to create an automated optical inspection (AOI) embedded detection system for the ornamental fish, which has successfully applied artificial intelligence technology to accelerate ornamental fish inventory and length information digitization.

The AOI embedded detection system for the ornamental fish(Fig.1) uses a dedicated visual measuring mechanism and an artificial intelligence model of fish characteristics, such as texture, contour, color, etc., which can accurately calculate the fish quantity and length information with an accuracy rate of 95% (approximately 100 per measurement). The system support CWG to digitize the quantity and size of each import and export goods quickly, so that the quality is guaranteed. The important achievement in the digitalization of ornamental fish, which is established by this system, can be further integrated with relevant information of suppliers such as fishery information, species and production lists, production records and inventory (Fig.2). These data can be integrated to calculate the production capacity in Taiwan and be shared with international buyers. Finally, through the big data analysis, the suppliers are able to get the sales information, such as hot-selling items, quantity demand, regional differences and market sales trends in the off or peak seasons, which can help them to plan production better.

After importing the results of this project, CWG will be able to reduce the labor costs, establish digital information and improve product quality. In addition, CWG has fully grasped the suppliers’ capacity information, which will provides much accurate sales information to customers and establishes excellent credit services. In the future, we will be able to use big data to predict global market trend and report back to suppliers for optimize breeding schedules as well as increasing the production capacity of Taiwan’s ornamental fish industry.