Realization of tree and grass recognition based on AlexNet

With the development of artificial intelligence technology, China is playing an increasingly important role in the field of object neural network recognition. In the past, people's understanding of neural network recognition was limited to the template matching model, which was simple and clear, but the model emphasized that the image must be completely consistent with the template in the brain to be recognized. However, AI recognition should not only recognize images consistent with the template, but also recognize images inconsistent with the template. The accuracy of the final training results reached 99.15%, exceeding the existing expectations. And AlexNet was born in 2012. On the model, AlexNet contains several relatively new technical points, and has successfully applied ReLU, Dropout, LRN and other Tricks in CNN for the first time. At the same time, AlexNet also uses the GPU for computing acceleration.