A Prototype System to Count Nursery Pacific White Shrimp using Image Processing

摘要

This paper proposes a counting system for estimating the biomass of nursery-stage pacific white shrimp in an indoor cultured tank using an image processing technique. The outcome is to reduce an over feed and water quality issue. As a result, it leads to the reduction of an operation cost in an indoor production system because the operation cost is mainly caused by feed and water treatment. Shrimp's biomass is directly used to determine an amount of feed. In Thailand, farmers currently use the sampling cup to sample and estimate a shrimp density. This traditional method depends on the experience and provide roughly estimation. Moreover, farmers are facing a labor issue. The proposed system is designed to estimate a shrimp's density automatically. The system consists of a sampling cup, sampling mechanism and counting unit. In image processing, colour, size and shape are used as three main features of the nursery-stage shrimp for distinguishing from their environments. The proposed method leads to more precise estimation of the density. The experimental results obviously reveal the effectiveness of the proposed system.