摘要
针对织物纹理在人工视觉分类中存在的分类准确率不高,效率低的问题,给出一种基于AdaBoost算法的色织物纹理图像分类方法.该方法首先对采集的织物图像进行人工标记;然后,提取所有图像的局部二值模式(local binary pattern,LBP)特征建立训练集;最后,基于AdaBoost算法对提取的色织物特征学习分类模型实现纹理图像的分类.实验结果表明,选择LBP28算子和弱分类器的个数为30时,基于AdaBoost分类算法对复杂色织物的纹理分类具有较高的识别率,其中对于梭织类和斜纹类织物分类准确率可达到100%,证实此方法的有效性.
To solve the low classification accuracy and low efficiency problem of color fabric in the human visual classification,a method which is based on the AdaBoost local binary pattern feature to classify the color fabric is proposed.Firstly,these collected fabric images are manually marked.Then,the local binary pattern features of images are extracted and a feature set is created.Finally,the extracted features are trained to form a model,tested and classified based on AdaBoost.The experimental results show that when the LBP8^2 operators are chosen and the number of weak classification is30,the AdaBoost algorithm can achieve a higher recognition rate for the texture classification of complex color fabric,and the classification accuracy of woven fabric and twill fabric can reach100%,confirming the effectiveness of the proposed method.
作者
李鹏飞
闫亚娣
张凯兵
王珍
朱丹妮
LI Pengfei;YAN Yadi;ZHANG Kaibing;WANG Zhen;ZHU Danni(School of Electronics and Information, Xi′an Polytechnic University, Xi′an 710048,China)
出处
《西安工程大学学报》
CAS
2018年第6期670-677,共8页
Journal of Xi’an Polytechnic University
基金
国家自然科学基金面上项目(61471161)
西安工程大学博士科研启动基金(BS1616)
