基于深度学习网络的智能养殖AI算法服务平台构建方法

Construction method of intelligent farming AI algorithm service platform based on deep learning network

为将传统农业与现代科技实现有效结合,实现养殖过程的智能化,构建一种基于深度学习网络的智能养殖AI算法服务平台。该平台包括基于Yolox模型目标检测算法、基于Mask R-CNN模型的目标分割算法和虚拟化资源池,前两者分别实现牲畜点数和测长估重,后者实现完善硬件的调度支撑方法。这样的处理方式可以加快算法收敛速度、降低参数计算量、提升系统性能,适用于畜牧业中的图像处理任务,实现对牲畜的全面管理和监测。

In order to realize an effective combination of traditional agriculture and modern science and technology,and to realize the intelligence of the farming process,a deep learning network-based intelligent farming AI algorithm service platform is constructed.The platform includes a target detection algorithm based on the Yolox model,a target segmentation algorithm based on the Mask R-CNN model,and a virtualized resource pool,the first two of which implement livestock point counting and length measurement and weight esti⁃mation,respectively,and the latter implements a scheduling support method for perfecting hardware.Such a processing method can accel⁃erate the convergence speed of the algorithm,reduce the amount of parameter computation,and improve the system performance,which is suitable for image processing tasks in animal husbandry to realize the comprehensive management and monitoring of livestock.