The trading and quality control of fries in the high-value aquaculture industry, such as giant freshwater prawns, lobsters, and ornamental fish, require significant labor and personal experience. During transactions, the process of counting small fries involves eliminating dead fries and avoiding errors from foreign objects (algae, small debris) or debris from the fries themselves (a shell or part of a dead fry), which is a frequent challenge. This study introduces an approach that employs both hardware and software solutions. Our solution includes calculating and selecting a small container with an appropriate lighting system to ensure stable lighting conditions for fries in water. This setup helps produce digital images from the camera that are consistent and with clear contrast between details in the image. Alongside this, we developed a counting tool that uses the YOLO model to classify fries from common debris, providing accurate fry counts with shorter processing times compared to manual labor. The sample used for this article involves small fries of shrimp and fish species commonly found in the market.