LAUSR

Microinjection of zebrafish larvae is one of the most common procedures in biomedical experiments. Majority of microinjection work of zebrafish larvae in the laboratory is operated manually. Manual operation is influenced by many factors, and it is difficult to maintain steady efficiency and high success rate. To solve this problem, a novel automated robotic system is proposed for batch microinjection of zebrafish larvae in this paper. For immobilizing the zebrafish larvae in batches with standard posture, an agarose supporting base is specially designed in a Petri dish. It is fabricated to hold the larvae in different orientations, which facilitates a rapid alignment of the randomly placed larvae. To accurately identify the injection targets of zebrafish larvae, a machine vision scheme based on deep learning is introduced. A prototype microinjection system is developed by integrating the hardware equipment and machine vision software, which is able to complete the injection operation quickly and accurately. The experimental results reveal that the system provides higher injection success rate, better stability, and better performance than those of existing microinjection approaches.