基于迁移网络的长江上游保护鱼类识别方法研究

Identification of Protected Fish in the Upper Reaches of the Yangtze River Based on Transfer Network

长江珍稀鱼类种群濒危程度日趋加剧,加强渔业从业人员保护意识并为其提供有效的保护鱼类参考是非常必要的.提出了一种基于迁移学习的网络模型,该模型无需大规模数据集驱动且适用于低算力设备,对于识别长江上游保护鱼类任务而言,它可以实时扩充数据集并以较快的速度进行训练且达到较高精度,对于普通手机用户而言,有较高的实用价值.

The endangered degree of rare fish in the Yangtze River is becoming more and more serious.It is necessary to strengthen the protection awareness of fishery practitioners and provide effective reference for fish protection.A network model based on transfer learning is proposed.The model does not need large-scale data set drive,and is suitable for low computing power devices.For the task of identifying fish protection in the upper reaches of the Yangtze River,it can expand the data set in real time and train almost in real time with high precision.It has high practical value for ordinary mobile phone users.