Multi-targets Object Mine Rock Image Processing And Analysis Online

In mining processing, in order to analyze the characteristic parameters of mine rocks precisely, it is necessary to segment rock fragment image accurately in on-line monitoring system. To meet the demand, the paper focuses on the rock fragment images threshold segmentation algorithms, and achieves the acquisition, processing and analysis of rock fragment images with image processing technology.The main contents of this paper include image pre-processing, image segmentation and parameter measurement of rock fragment images. Firstly, the acquisition, pre-processing and image segmentation of rock fragment images are introduced. Then aiming at touched rock regions which are uneasy to separate, the fuzzy comprehensive evaluation method is used for estimating the characteristic parameters of rock regions and determining whether the rock region needs a further segmentation based on the estimation. Finally, the automatic segmentation, boundary tracking and feature extraction for rock fragment images have been achieved. The application of the presented algorithm can meet the demand of mine on-line monitoring system. The main contents of this paper can be concluded to the following points:1、Image Pre-processingIn the processing of rock image acquisition, large numbers of noise in the external environment may impact the follow-up work, so it is necessary to pre-process the original images. The paper firstly completes the gray-scale transformation of color image, and then finishes smoothing to weaken the impact of noise.2、Rock fragment image Thresholding SegmentationAfter completing the pre-processing, we need to segment the rock fragment images. In order to meet the requirements of the online system, we choose the threshold segmentation. The paper compares several threshold segmentation algorithms and implements them, and chooses a threshold method which is suit for rock fragment images segmentation.3、Rock Regions CharacteristicsOn the basis of the initial segmentation, various rock regions information can be achieved includes shape information, edge information and texture information. All the information is the basis for the next comprehensive fuzzy evaluation. In the process of acquisition, the position of a region should be determined at first, and then the shape information can be acquired, at last the edge and texture information has been acquired in the corresponding position of the gray-scale image.4、Fuzzy Comprehensive EvaluationAccording to information of each rock region, the evaluation of the different regions based on fuzzy comprehensive evaluation technique has been completed to determine whether the rock needs further segmentation or not. The factor set consists of the region information, and the weight set is determined by the status of the region information. The fuzzy memberships function of each parameter is determined through statistical methods and then the final evaluation outcomes.5.Rock fragment image segmentation results analysisThe final segmentation algorithm is based on fuzzy comprehensive evaluation of rock fragment images. And then distribution of rock fragmentation comes out base on the segmentation results. A comparison has been made among the traditional algorithm and the final algorithm, and the rock distribution analysis has been achieved on the bsais of the final results.