Cow Individual Identification Based on Convolutional Neural Network

With the advent of the artificial intelligence era, the intelligent monitoring video analysis technology is deeply applied to the management decision of livestock and poultry farming, and the individual identification of dairy cows can improve the efficiency of video analysis. In this paper, an individual identification method based on convolutional neural network is proposed for the automation and precision of dairy cows. Using the video capture device to obtain the 360° torso information of the cow, the cow's video is obtained by frame acquisition to obtain the cow's torso image information, and the residual learning deconvolution neural network is used to denoise the 21600 cow images. Randomly selects pictures to define the training set and testing set as the ratio of 9:1 and selects the InceptionV3 network to train the main network of the individual identification of the cows. Through the cross-validation of the training set and the test set, the top_1 recognition accuracy rate of the single image can reach 87%, and the top_3 recognition accuracy rate is 93%.Through the improvement of InceptionV3 neural network, the top_1 recognition accuracy of multiple images can reach 92%, and the recognition accuracy of top_3 is 95%.