Oyster aquaculture plays a vital role in the Gulf of Mexico’s economy and culture. However, the region’s shortage of oyster hatcheries and nurseries poses a risk to the industry’s sustainability, as farmers may face unreliable access to oyster seed (juvenile oysters). A potential solution is for farmers to establish their own nurseries and raise seed. This project aimed to facilitate farmer-run nurseries by advancing quality management techniques. Accurate recordkeeping is essential for effective quality management in oyster nurseries. In an oyster nursery, records of water quality, the number of seed, and seed growth and survival rates must be kept. In upweller systems, where seed grow quickly, overcrowding is a concern, so careful density measurements are crucial. Farmers estimate seed densities and counts by weighing a subsample and calculating the total seed count and density based on the sample’s weight. Counting seed daily is often too time-consuming, and results can vary by operator. To address this, a cost-effective quality management package consisting of software and hardware components was developed.
The software component consists of image analysis software that automatically counts the number of seed in an image (Figure 1). This provides an alternative counting method that does not require a microscope and allows seed counts to be more time-efficient. The software was coded in Python (v3.9.13, Python Software Foundation, 2024). When using the software, the time required for the seed-counting step was 65% faster than manually counting. Furthermore, the total time to perform all seed counting steps decreased by 10% when using the software compared to manual counting. The hardware component is in development and will include a series of scoops of different volumes to collect seed in different size classes. These collection tools allow farmers to quickly transfer a known seed volume into the second device type, a counting chamber. The counting chamber will be made of material that allows for high-contrast images of the seed, to facilitate better software results. In addition, the chamber will have an imprinted grid to provide a reference for manual seed-size estimations and will be able to accommodate a range of size classes.