To estimate spawning time in shrimp, we use infrared underwater photographya trained AI model to monitor gonad maturity in shrimp broodstock.

In the darkened conditions used for shrimp gonad development, researchers usually use flashlights to visually determine gonad maturity,then use their best judgment to estimate the spawning time. However, this time-consuming process disturbs the highly sensitive shrimp,errors are easily made – which makes it an attractive candidate for an alternative, AI-based solution. Our novel technology accounts for the shrimp pond environment, the absence of any light sourcethe behavior of the shrimp: we use underwater infrared photography to record subtle changes in the broodstock gonads without disturbing the shrimp, after which trainingdata analysis is used to intelligently judge the spawning time of each shrimp.

Currently, most of the global shrimp farming industry still uses manual inspection methods to estimate shrimp gonad maturityspawning times. But even professional researchers often make misjudgments,the introduction of light into the darkened broodstock environment can easily cause the frightened shrimp to retract its gonads, after which they will require more time to ripen again. By contrast, our unique AIoT system combines infrared underwater photography, image enhancementrecognition with AI modelingdata analysis to provide both improved accuracya gentle, disturbance-free method. This kind of shrimp-friendly technology represents an important step forward for equipment automation in the aquaculture sector.