基于卷积神经网络的树叶识别的算法的研究

Based on Leaves Convolutional Neural Network Recognition Algorithm

该文研究了将卷积神经网络应用在树叶识别方面,并通过卷积过程对图片进行可视化。实验表明,卷积神经网络应用在树叶识别达到了92%的识别率。另外,将此神经网络与支持向量机进行比较研究,从试验中可以得出,卷积神经网络在无论是精度方面还是速度方面都要优于支持向量机,可见,卷积神经网络在树叶识别方面具有很好的应用前景。

In this paper,the convolution neural network recognition in the leaves,and the process by convolution of image visualization.Experiments show that the neural network application identification convolution leaves a 92% recognition rate.In addition,this neural network and support vector machine comparative study can be drawn from the study,convolutional neural network in either speed or accuracy better than support vector machines,visible,convolution neural network in the leaves aspect has good application prospects.