基于多颜色分量CLBP提取的浮选泡沫状态识别

针对不同状态的浮选泡沫图像之间纹理结构相似、颜色差异不明显的问题,提出一种基于色调、饱和度和亮度(HSV)颜色空间的完全局部二进制模式(CLBP)纹理提取的浮选泡沫状态识别方法。首先使用双域去噪在保留纹理细节的同时滤除图像噪声;然后转换为HSV图像,在H、S和V颜色分量上分别提取三个尺度的CLBP纹理特征,将提取的纹理特征归一化后线性排列,建立高维度的纹理分类模型;最后通过一对一模式的支持向量机分类器对四类泡沫状态的样本集进行纹理提取后的分类训练与测试。结果表明,该方法对不同浮选泡沫状态的分类正确率较高,优于其他纹理描述方法,适用于浮选泡沫状态的识别。

Flotation bubble state recognition based on multi-color component CLBP extraction

Aiming at the problem that the texture structure of the flotation foam images in different states is similar and the color difference is not obvious, a method based on Hue, Saturation, Value (HSV) color space for Completed Local Binary Pattern (CLBP) texture extraction is proposed. First, the Dual-domain image denoising is used to filter the image noise while preserving the texture details, and then converted to HSV images, and the three-scale CLBP texture features are extracted on the H, S, and V color components respectively. The extracted texture features are normalized and linearly arranged to establish a high-dimensional texture classification model. Finally, through the one-versus-one support vector machine classifier, the four-class bubble state sample set is subjected to texture extraction and classification training and testing. The results show that the method has higher classification accuracy for different flotation foam states, and is superior to other texture extraction methods, which is suitable for the identification of flotation foam states.