News
:::

The Intelligent Image Recognition System for Assisting Paddy Seed Inspection

The Intelligent Image Recognition System for Assisting Paddy Seed Inspection

  “The peanut was selected by a computer” is a famous slogan of sweet. The producer selected the peanut by a computer to control the product quality. Now the computer can select not only peanut but also the paddy seeds! The latest research of Taiwan Seed and Seedling Improvement and Propagation Station (TSIPS) shows the usage of image recognition system on paddy seeds to check the quality.

  Rice is the main crop of Taiwan, the common varieties are included ‘Tainan No. 11’, ‘Taiken No. 9’, ‘Tainung No.71’ and ‘Taichung Sen No. 10’ etc. The “Seed Multiplication and Certification System” takes an important rule on keeping the paddy seed quality and the Seed Testing Laboratory (STL) of TSIPS is responsible for the seed testing task of Taiwan. The seed moisture content, purity and germination of Foundation Seed, Stock Seed and Extension Seed were qualified by STL. Since the various species and ‘Black Rice’ were widely found in paddy seed multiplication fields, the seed testing tasks for ensuring the seed quality are quite fundamental.

  The purity and germination tests now are only manual inspecting the grain shape and seedling structure that conducted by well-trained inspectors and are time and energy consuming. The results may also effected by fatigue. The Intelligent Image Recognition Technology that researched and developed by TSIPS established the digitized the paddy grain shape and reached 90% recognizability of 7 main rice varieties including ‘Tainan No. 11’, ‘Taiken No. 9’ , ‘Taiken No. 14’, ‘Taichung No. 192’, ‘Taichung Sen No. 10’, ‘Tainung No.71’ and ‘Taoyuan No.3’. The following research will focus on establishment of seedling images capture, integration of scanner and container; build the seedling image database and image recognition algorithm of ‘Black Rice’. This research is expected to reduce the seed testing manpower, improve the inspection efficiency and decrease the manual mistakes. Therefore, this system also stabilizes the supply of high quality of domestic paddy seed.

Fig. The seedling image that captured by intelligent image recognition system.

Fig. The seedling image that captured by intelligent image recognition system.