复杂混合浮选泡沫图像分割方法研究

针对浮选泡沫图像难以直接识别的问题,提出了一种基于形态学的分水岭改进算法,利用CLAHE、最小最大值滤波对泡沫图像进行预处理,采用形态学方法进行图像重建,最后利用分水岭算法对重建后的图像进行分割。仿真结果表明,该算法可有效地对复杂混合泡沫图像进行分割。     

基于Contourlet变换的模糊聚类和膨胀重构相结合的浮选图像增强方法

针对用分水岭算法分割铅锌矿泡沫图像时标识点难以提取及气泡边界模糊问题,提出了模糊C均值聚类和二值开重构相结合的方法及基于Contourlet变换的泡沫图像增强方法。首先对图像进行聚类,提取出灰度值较大的一类点,对提取出获得的图像用不同的结构元素进行二值开重构操作除掉一些灰度值较大的边缘点和噪声,得到气泡内部高亮点作为标识点。然后对泡沫图像做Contourlet变换提取高频系数,构造自适应增强函数作用于高频系数,利用增强后的高频系数做Contourlet反变换以增强泡沫图像。最后在上面获得的标识图像和增强图像的基础上用分水岭变换对图像进行分割。试验结果表明,该方法能够对气泡边界模糊的铅锌矿图像实行有效分割。     

煤泥浮选泡沫数字图像处理中的边缘检测算法研究

通过分析煤泥浮选泡沫图像的特点和应用常规边缘检测算法所存在的问题, 提出了综合应用煤泥浮选泡沫图像灰度和梯度信息进行煤泥浮选状态监控的观点, 介绍了使用基于Sobel算子改进的方向算子对煤泥浮选泡沫图像进行处理的方法。     

高炉渣离心粒化图像颗粒分割研究

高炉渣机械离心粒化过程中,基于图像识别的颗粒分割是实现粒径实时检测的关键因素。针对于高炉渣图像颗粒黏结而给测试带来困难的问题,在对图像做去噪处理后,结合改进的分水岭算法做颗粒分割,实现了黏结颗粒的准确分割。针对分水岭算法缺陷,结合积分图像的快速算法,对颗粒图像做局部阈值的二值化处理,并由形态学做细节弥补,采用了基于形态学重构距离图像算法以改进分水岭算法来分割颗粒图像,提高了处理效率,同时抑制了过分割现象,并准确地分割了黏结颗粒的图像,该算法为准确快速提取颗粒的特征参数奠定了基础。     

煤泥浮选泡沫图像分割算法研究

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。     

基于控制标记符分水岭的重叠细胞分割

医学上存在细胞之间有重叠的现象,可以利用图像处理的方法解决。由于传统的图像处理法—分水岭算法容易产生过分割和对噪声敏感的问题,所以,本文首先对图像计算梯度,然后对目标和背景进行标记,最后根据被标记的二值图像进行修正,通过分水岭算分进行分割。实验证明,该方法能很好地分割重叠细胞,并给出其MATLAB程序。     

煤泥浮选泡沫图像分割与特征提取

针对在煤泥浮选泡沫图像中的煤泥气泡互相粘连、边界模糊的情况，根据分水岭变换模拟浸水原理，提出一种新的、有效的分割算法，同时应用数学形态学中的腐蚀和膨胀算法，解决了经典分水岭算法不能处理的过分割和欠分割问题，使得粘连的煤泥气泡得到了有效的分割.然后计算出各个煤泥气泡的横截面积、周长、形状等物理特征参数，这些参数是实现自动控制浮选过程的重要依据.     

煤泥浮选气泡比表面积的计算方法

为了从基于分水岭变换分割后的煤泥浮选泡沫图像上获得更加准确的气泡大小的分布情况，需要计算气泡的比表面积.由于气泡的不规则性，使用等效椭圆比等效圆来计算面积、体积更能反应真实的气泡情况.通过对比表面积进行不同的加权，比较表明，使用气泡投影的横截面积作为加权系数是描述气泡分布情况最适合的方法.同时还讨论了气泡特征与浮选操作条件之间的关系，为计算机在线控制浮选过程建立基础.     

基于U-Net和改进分水岭算法的露天矿爆堆矿石图像分割方法

爆破大块率是反映爆破效果的关键指标,其统计精度主要依赖于爆堆矿石图像分割的准确性。由于爆堆矿石图像中存在矿石目标分布密集、边缘对比度低等问题,致使传统图像分割方法难以准确分割爆堆矿石图像。因此,提出了一种基于U-Net和改进分水岭算法的露天矿爆堆矿石图像分割方法,以实现大块率的精准统计。首先利用无人机在哑巴岭露天矿爆破现场拍摄爆堆矿石图像,制作爆堆矿石图像数据集;然后利用深度学习算法建立了UNet网络架构,同时融合了高级语义信息和低级语义信息,建立了爆堆矿石图像分割模型,再利用训练后的模型对爆堆矿石图像进行初步分割,进一步采用基于距离运算的分水岭算法优化了分割结果;最后评估了该方法的分割精度。试验结果表明:该算法可准确分割露天矿爆堆矿石图像,为露天矿爆破大块率统计、爆破效果智能评价提供技术支持。     

基于灰色系统理论的煤泥浮选泡沫数字图像处理算法研究

探讨了灰色系统理论在边缘检测中的应用,分析了应用常规边缘检测算法进行煤泥浮选泡沫图像处理所存在的问题,提出了在煤泥浮选状态机器视觉监控中应用基于灰色关联度计算进行图像边缘检测的观点,介绍了应用基于灰色关联度计算对煤泥浮选泡沫图像进行处理的方法。     

Froth delineation based on image classification

This paper describes a set of image segmentation algorithms for mineral froth images, based on gray-value valley detection and a kind of image classification. The size, shape, texture and color of froth bubbles are very important pieces of information for production optimization in mineral processing. In order to determine these parameters, bubbles in a froth image first have to be delineated. Froth images display a large variation of image patterns and quality, thus it is difficult to use only a single algorithm for segmenting all images. To achieve successful segmentation the images are first classified into image classes. Then sets of segmentation algorithms are used, based on the different image classes. The segmentation algorithms and classification algorithms have been tested in a laboratory and in industrial on-line systems for froth images, the test results show that they are robust for froth images. The processing speed for the segmentation algorithm is much faster than for a standard morphological segmentation algorithm. The processing accuracy is comparable to manual drawn result. This test shows that the algorithms work satisfactorily.     

浮选机矿浆相气泡图像采集系统的设计

在浮选优化控制领域，有关泡沫图像采集、处理的研究已被广泛应用。现有的研究绝大多数关注的是采集自浮选机泡沫相表面的图像数据，而对于矿浆相，由于液体本身环境恶劣，透光性极差，具有腐蚀性等因素，内部气泡图像难以获取，从而影响技术发展。针对此问题，文本设计了一种新的气泡图像采集系统，能够直接在矿浆溶液里原位采集图像数据，并提出利用深度学习语义分割模型实现矿浆相气泡图像的精确分割。实验结果表明，该系统在实验室环境下能够采集到清晰的矿浆相气泡图像，基于小样本利用预训练模型进行迁移学习，验证集的气泡分割精度达到0.943，具有重要的研究和应用价值。     

The froth stability column: Measuring froth stability at an industrial scale

Froth structure and stability are known to play important roles in determining mineral flotation recovery and selectivity. However, measuring froth stability in a consistent manner remains a significant challenge, especially at an industrial scale. Following preliminary tests on a copper concentration plant, a quantitative dynamic stability measure is investigated in this study in order to extend the results over a wider range of conditions and on a different ore body. The technique is based on the Bikerman foam test and uses a non-overflowing froth column to quantify froth stability.Experiments were carried out using an automated version of the froth stability column under different operating conditions. Air flowrate was the key operating variable. Tests were reproduced on a single flotation cell of a Platinum Group Metals concentrator. The froth stability factor, β, was measured for each operating condition, and compared with the air recovery in the cell, α, which was measured using image analysis. The froth stability column results gave the same trends as image analysis. In particular the froth stability factor was found to be linearly related to the actual fraction of air overflowing the cell.The metallurgical results clearly indicated that changes in air flowrate result in variations in flotation performance that can be attributed to changes in froth stability. The results showed that high froth stability conditions occur at intermediate air flowrates, and result in improved flotation performance. It is found that the froth stability column is a simple, cost-effective and reliable method for quantifying froth stability, and for indicating changes in flotation performance.     

煤泥浮选泡沫图像灰度直方图及其统计纹理特征研究

针对煤泥浮选泡沫图像灰度直方图特征问题，进行了实验室浮选柱试验，采集了50组煤泥浮选泡沫图像，分析了泡沫图像及其直方图的类别及形态，提取了浮选泡沫图像直方图统计纹理特征参数，研究了泡沫直方图统计纹理特征参数随浮选时间的变化关系。研究表明，统计纹理特征参数-方差能够表征浮选泡沫图像直方图特征，并与特定的泡沫状态相关，可为煤泥浮选视觉监控系统提供泡沫状态信息。     

图像处理技术在浮选过程中的应用

综述浮选泡沫处理参数的算法和图像处理技术在浮选过程控制中的应用。比较图像处理技术与其它浮选过程检测控制技术的优缺点,指出泡沫图像处理技术仪表化的技术前景。     

基于机器视觉及IHS-LSSVM的浮选回收率预测

针对矿物浮选过程中回收率难以在线检测的问题,提出了一种基于机器视觉技术、改进和声搜索算法(HS)优化LSSVM参数的预测方法(IHS-LSSVM)。首先提取浮选泡沫图像的颜色、气泡尺寸、分形系数、运动速度等多个特征,接着改进和声搜索算法的"调音"策略,进而找到LSSVM的最优参数,并建立浮选回收率预测模型,最后采用铅矿浮选现场数据对模型进行测试。试验结果表明,相对于现有方法,该预测方法提高了预测精确度。     

Recognition of the operational statuses of reagent addition using dynamic bubble size distribution in copper flotation process

The output PDF (probability density function) shape of surface bubble size in froth flotation is believed to be closely related to the operational statuses of reagent additions. An online operational status recognition method for quality evaluation of reagent addition is presented based on adaptive learning of the dynamic distribution features of the surface bubble size. We avoid the bubble over-segmentation problem by exploring an improved image segmentation algorithm to get the accurate bubble size statistics taking account of the local regional distribution of the image brightness value. By utilizing the kernel density estimation, we obtain the PDF and CDF (cumulative distribution function) of the bubble size statistics effectively. The distribution features of the bubble size statistics under the PDROS (pre-defined reagent operation statuses) are learned by FNC (furthest neighbor clustering), successively, the current health status of the reagent addition in the test time period is inferred by Bayesian inference according to the dynamic change of the bubble size PDFs; what is more, the statistical distribution features under PDROS are updated online according to the disturbance of the process operation conditions. This status recognition method is tested and practically applied in a copper ore beneficiation plant. The experimental results on the real production data demonstrate that this method outperforms other quasi machine vision based production condition recognition methods with much lower error recognition rate. It paves the way for the realization of an optimal control and fault diagnosis for reagent addition in the flotation process operation.     

基于浮选泡沫图像特征参数的应用研究

基于泡沫图像分析仪在大山选厂的具体应用,研究浮选泡沫的物理意义以及获取的浮选泡沫图像具体特征参数,提取100个矿浆样以及相对应的泡沫图像样,建立两者的数学模型,用该模型对铜精矿品位进行实时预测,有效地指导了现场的实际生产操作。