In aquaculture breeding or production programmes, counting juvenile fish represents a considerable cost in terms of the human hours needed. In this study, we explored the use of two state‐of‐the‐art machine learning architectures (Single Shot Detection, hereafter SSD and Faster Regions with convolutional neural networks, hereafter Faster R‐CNN) to augment a manual image‐based juvenile fish counting method for the Australasian snapper ( Chrysophrys auratus ) bred at The New Zealand Institute for Plant and Food Research Limited. We tested model accuracy after tuning for confidence thresholds and non‐maximal suppression overlap parameters, and implementing a bias correction using a Poisson regression model. Validation of image data showed that after tuning, bias‐corrected SSD and Faster R‐CNN models had mean absolute percent errors (MAPE) of less than 10%, with SSD having MAPE of less than 5%. Comparison of the results with those from manual counts showed that, while manual counts are slightly more accurate (MAPE = 1.56), the machine learning methods allow for more rapid assessment of counts and thus facilitating a higher throughput. This work represents a first step for deploying machine learning applications to an existing real‐life aquaculture scenario and provides a useful starting point for further developments, such as real‐time counting of fish or collecting additional phenotypic data from the source images.
Several lepidopterans are pests in horticulture and pose biosecurity risks to trading countries worldwide. Efficient species-specific semiochemical lures are available for some of these pests, facilitating the implementation of surveillance programmes via trapping networks. These networks have a long history of success in detecting incursions of invasive species; however, their reliance on manual trap inspections makes these surveillance programmes expensive to run. Novel smart traps integrating sensor technology are being developed to detect insects automatically but are so far limited to expensive camera-based sensors or optoelectronic sensors for fast-moving insects. Here, we present the development of an optoelectronic sensor adapted to a delta-type trap to record the low wing-beat frequencies of Lepidoptera, and remotely send real-time digital detection via wireless communication. These new smart traps, combined with machine-learning algorithms, can further facilitate diagnostics via species identification through biometrics. Our laboratory and field trials have shown that moths flying in/out of the trap can be detected automatically before visual trap catch, thus improving early detection. The deployment of smart sensor traps for biosecurity will significantly reduce the cost of labour by directing trap visits to the locations of insect detection, thereby supporting a sustainable and low-carbon surveillance system.