基于改进神经网络算法的植物叶片图像识别研究

Plant Leaf Image Recognition Based on Improved Neural Network Algorithm

为了提高植物叶片图像的识别率,采用改进神经网络算法,通过径向基函数神经网络建立模型;采用多环量子算法确定各环量子个体选择概率,量子旋转门在一定范围内动态调整,不同环上节点信息共享概率非线性动态变化;对植物叶片图像进行识别,包括形状特征、纹理特征;通过多环量子算法实现径向基函数神经网络参数寻优。实验结果表明,本文算法对植物叶片图像的几何特征、纹理特征、综合特征的平均识别率分别为91%,89%,93%,与其他算法相比较高,训练、识别时间分别为3.5s、2.5s。

In order to increase the recognition rate of plant leaf images, the improved neural network algorithm is proposed. The model is established by radial basis function neural network. The multi loop quantum algorithm is used to determine the selection probability of each quantum individual, and the quantum rotation gate is dynamically adjusted in a certain range, and the node information of different rings shares the probability of nonlinear dynamic changes. The plant leaf image recognition includes shape features and texture features. The multi loop quantum algorithm is used to realize the radial basis function neural network parameter optimization. The experimental results show that the proposed algorithm has a higher average recognition rate of plant leaf image than other algorithms, with the geometric features 91%, texture features 89% and comprehensive features 93%, and the training and recognition time are 3.5 s and 2.5 s respectively.