高密度工厂化水产养殖鱼种识别模型

通过高清摄像头捕捉鱼群图像，模型快速提取形态特征（体型、体色、鳍条等），比对海量鱼种数据库，1快速完成识别，准确率超 98%。​
可实时区分混养品种、监测苗种纯度，辅助优化养殖密度与投喂方案。减少人工识别误差，降低苗种错养风险，助力提升单产 10% 以上，为集约化养殖提质增效。

This fish species recognition system captures images of fish schools via high-definition cameras, rapidly extracts morphological features such as body shape, body color, fin rays and other relevant traits, compares against a large-scale fish species database, and achieves rapid identification with an accuracy rate exceeding 98%. It can distinguish polycultured fish varieties in real time, monitor the purity of fish juveniles, and assist in optimizing stocking density and feeding regimes. It reduces manual identification errors, lowers the risk of misbreeding of fish juveniles, helps boost per-unit yield by over 10%, and improves quality and efficiency for intensive aquaculture.