基于叶片图像算法的植物种类识别方法研究

发布时间：2018-03-15 01:01

  本文选题：叶片　切入点：植物种类　出处：《浙江农业学报》2017年12期 　论文类型：期刊论文

【摘要】：为了提高植物种类的识别率,采用叶片图像算法。首先建立植物种类特征模型,包括植物叶片颜色特征、形状特征、纹理特征;然后确定径向基函数神经网络的输入层、输出层、隐含层之间的关系;接着对径向基函数个数、中心及宽度优化,基于梯度下降方法对权重参数计算,自适应调节学习率;最后给出了植物种类识别过程。实验仿真选择植物叶片颜色特征、形状特征、纹理特征的特征量分别为6、7、7个,其中本文算法对植物种类识别的三个组合特征平均识别率为93.5%,高于单个特征、两个组合特征的平均识别率,形状特征对识别率所起的作用最大。
[Abstract]:In order to improve the recognition rate of plant species, the leaf image algorithm is used. Firstly, the plant species feature model is established, including the color feature, shape feature, texture feature of plant leaf, and then the input layer of radial basis function neural network is determined. The relationship between the output layer and the hidden layer, then the number, center and width of the radial basis function are optimized, the weight parameters are calculated based on gradient descent method, and the learning rate is adjusted adaptively. Finally, the process of plant species recognition is given. The number of color features, shape features and texture features of plant leaves are 6 7 and 7 respectively. The average recognition rate of the three combined features of this algorithm is 93.5, which is higher than that of a single feature. The average recognition rate of the two combined features and the shape feature play the most important role in the recognition rate.
【作者单位】： 黄河水利职业技术学院;
【基金】：基金项目:中国国家专利(公开号CN202189701U) 河南省科学技术成果(豫科鉴委字2013年第201号)
【分类号】：Q94;TP391.41
，

本文编号：1613705