摘要
图像特征提取是图像处理重要的研究内容之一,针对树皮图像,提出了一种基于分形理论的树皮图像特征提取方法。首先将树皮RGB彩图转换成灰度图,对灰度图进行局域分形特征提取,得到分形特征灰度图,再计算此特征图的水平和垂直两个方向的像素数值之和;然后用此特征和标签进行支持向量机(SVM)训练和测试,得到样本的分类结果。试验结果表明,通过设置不同的特征长度,提出的方法能很好地提取树皮特征,识别率分别为92. 78%、93. 33%、93. 88%。
Image feature extraction is one of the important research contents of image processing,an image feature extraction method based on the fractal theory is proposed for bark images. Firstly,a bark RGB color image is converted into a grayscale image,and the localized fractal feature extraction is performed on the grayscale image to obtain a fractal feature grayscale image,and the sums of the pixel values in the horizontal and vertical directions of the feature image are calculate respectively. This feature and labels are used for support vector machine( SVM) training and testing to get the classification results of the samples. The experimental results show that the proposed method can extract the bark features well by setting different feature lengths,and the recognition rates are 92. 78%,93. 33% and93. 88% respectively.
引文
[1]张立军,李晓环.林业产业发展与森林资源保护分析[J].现代园艺,2018(12):114-118.
[2]赵作林.基于图像分析的北京地区杨树种类识别研究[D].北京:北京林业大学,2015.
[3]周熊,多化琼,王振柱,等.基于图像处理和SVM探索木材与树皮的新识别参数[J].西北林学院学报,2014,29(5):199-202.
[4]程玉柱,段一凡.基于PSO与水平集的桂花图像分割方法[J].中国农机化学报,2018,39(8):81-84+91.
[5]关雅卓.智能化识别技术在电子图像处理中的应用研究[J].中国管理信息化,2018,21(18):152-153.
[6]程玉柱,顾权,王众辉,等.基于深度学习的木材缺陷图像检测方法[J].林业机械与木工设备,2018,46(8):33-36.
[7]解志君,杨伟民.基于机器学习的红外激光图像特征定位技术[J].激光杂志,2018,39(9):96-100.
[8]王小龙,任亚群,葛海明,等.钙质结核土分形特性及表观特征研究[J].电力勘测设计,2018(S1):104-108.
[9]冯学晓,古险峰.基于分形理论的图像边缘特征提取算法[J].新乡学院学报,2018,35(6):40-43.
[10]王克奇,谢永华,陈立君.基于分形理论的木材纹理特征研究[J].林业机械与木工设备,2005(7):19-20.
[11]李可心,戚大伟,牟洪波,等.基于灰度共生矩阵与SOM神经网络的树皮纹理特征识别[J].森林工程,2017,33(3):24-27.
[12]刘泉菲,赵彬彬,周凯.基于分形维数的多尺度面目标匹配对相似性度量[J].长沙理工大学学报:自然科学版,2018(3):1-7.
[13]肖神清,周兴林,刘万康,等.沥青道路表面构造分布的多重分形特性[J].公路,2018,63(4):1-6.
