特征加权和模糊聚类算法的植物叶片识别研究

Plant leaf recognition based on feature weighting and fuzzy clustering algorithm

针对当前植物叶片识别正确率低、识别时间长等局限性,为了获得更好的植物叶片识别结果,提出了基于特征加权和模糊聚类算法的植物叶片识别方法。首先采用传感器对植物叶片图像进行采集,并对图像进行预处理,其次提取植物叶片识别特征,采用人工鱼群算法确定植物叶片识别特征权值,最后根据特征加权结果对植物叶片识别样本进行处理,通过模糊聚类算法实现植物叶片分类和识别,最在MATLAB 2017平台上的仿真测试。结果表明,文中方法可以正确识别各类植物叶片,植物叶片识别正确率高,误识率远远小于其它植物叶片识别方法,验证了文中方法的优越性。

In view of the limitations of low accuracy and long recognition time of plant leaf recognition,a method of plant leaf recognition is proposed based on feature weighting and fuzzy clustering algorithm to obtain better results of plant leaf recognition.Firstly,images of plant leaves are collected by senso and preprocessed to extract the original features of plant leaf recognition.The artificial fish swarm algorithm is used to determine the weight of plant leaf recognition features.Finally,the recognition samples of plant leaves are processed according to the weighted results of the features,and the classification and recognition of plant leaves are realized by using the fuzzy clustering algorithm.The simulation results on MATLAB 2017 platform show that this method can correctly identify all kinds of plant leaves,and the recognition accuracy of plant leaves is high,and the error rates are far less than those of other plant leaf recognition methods,which verify the superiority of this method.