煤泥浮选泡沫图像分割算法研究

Research on Image Segmentation Algorithm of Coal Flotation Froth

针对工业现场采集的泡沫图像噪声大、对比度低、气泡粘连等特点,提出了一种基于双边滤波与限制对比度自适应直方图均衡的分水岭分割模型。首先通过双边滤波对原始泡沫图像进行去噪处理;其次,针对泡沫图像对比度低的特点,采用限制对比度自适应直方图均衡(CLAHE)增强图像对比度以及防止噪声的放大,最后利用分水岭变换对预处理之后的图像进行分割处理。实验结果证明该模型的有效性,预处理之后分割的图像能够有效改善直接处理所带来的欠分割现象。

Froth image collected from industrial scene is usually characterized with the noise, low contrast and the adhesion of bubbles. In view of these, propose a segmentation model based on bilateral filtering and contrast limited adaptive histogram equalization. Firstly, the bilateral filtering was used to remove the noise of the original froth image. Secondly, the CLAHE was used to enhance the contrast and avoid the amplification of noise. Finally, the watershed transformation was employed to segment the image after process based the original watershed algorithm. Experimental results prove the effectiveness of the proposed model.