Sika Deer Behavior Recognition Based on Machine Vision

With the increasing intensive and large-scale development of sika deer breeding industry, it is crucial to assess health status by monitoring their behaviours. A machine vision–based method for behaviour recognition proposed in this paper. Google Inception Net (GoogLeNet) used optimise model First, number layers size were reduced. Then, 5 × convolution was changed two 3 convolutions, which reduced parameters increased nonlinearity model. kernel replace original extracting coarse-grained features improving model’s extraction ability. multi-scale module added enhance multi-faceted feature capability Simultaneously, Squeeze-and-Excitation Networks (SE-Net) included increase channel’s attention improve accuracy. The dataset’s images rotated reduce overfitting. For image rotation, angle multiplied 30° obtain dataset enhanced rotation operations 30°, 60°, 90°, 120° 150°. experimental results showed that rate 98.92%. Therefore, presented paper can be applied deer. will play an essential role promoting animal technology management.