【摘要】：為了準確地檢測出浮選過程中顆粒與氣泡是否發(fā)生碰撞,提出了一種基于形態(tài)學圖像處理和最大類間方差法(Otsu)的顆粒與氣泡碰撞檢測方法。首先利用圖像的預處理操作優(yōu)化圖像,通過最大類間方差法分割出目標區(qū)域與背景區(qū)域,然后結合最小二乘法將顆粒與氣泡擬合成圓,方便求取坐標信息及半徑,最后將所有的顆粒位置提取到一幅圖像中,實現了對顆粒位置的跟蹤,避免了連續(xù)幀圖像跟蹤的復雜程度。結果表明,該算法能夠快速地檢測出顆粒與氣泡是否發(fā)生碰撞,有效地節(jié)省了人工試驗的工作時間。
[Abstract]:In order to detect accurately whether particles collide with bubbles in flotation, a collision detection method based on morphological image processing and maximum inter-class variance method (Otsu) is proposed. Firstly, the image is optimized by image preprocessing, and the target area and background area are segmented by the method of maximum inter-class variance, and then the particle and bubble are combined with the least square method to obtain the coordinate information and radius. Finally, all the particle positions are extracted into a single image, and the tracking of the particle positions is realized, which avoids the complexity of continuous frame image tracking. The results show that the algorithm can quickly detect whether the particles collide with the bubble and save the working time of the artificial experiment.
【作者單位】： 華北理工大學電氣工程學院;華北理工大學礦業(yè)工程學院;
【基金】：國家自然科學基金項目資助(編號:51474087)
【分類號】：TD923;TP391.41
，

本文編號：2146906