Font Size: a A A

The Application Of Image Processing Techniques In Rock Blasting Fragmentation Analysis

Posted on:2013-05-10Degree:MasterType:Thesis
Country:ChinaCandidate:J HuFull Text:PDF
GTID:2268330401984783Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The size distribution of muck piles is analyzed in the existing literature so as tomeasure the ore blasting effect. Efficient and accurate identification of ore grain sizeis mine blasting optimization foundation, and it can bring huge economic benefits forthe mine production. The identification of ore particle size is an important researchtopic in today’s mining applications. The normal method is direct test and indirect testmethod. There are all kinds of defects in the direct test method, so the commonly usedmethod is indirect test method. In recent years, with the development of imageprocessing and pattern recognition techniques, single-picture photography methodbased on image processing technology become an important granularity statisticalanalysis method.The author summarizes the theory and analysis methods of the size distribution,and in-depth study the mine rock granularity image analysis applications. Differenttypes of minerals images are gathered for developing a statistical system based on theparticle size distribution of the rock image. The conditions of the open pit area is morecomplex, uneven illumination, form complex ore, the particle size distribution isuneven, so resulting in the gathered images containing a variety of noise, the unevendistribution of the gradation, the image detail is not clear. In this paper, an improvedrock image segmentation algorithm based on morphological reconstruction and tagextraction is presented.The core part of the pre-treatment is based on the two-stagegrayscale morphological reconstruction. The method makes a stronger image contrastand better smoothes the foreground and background, while there is no loss of edgeinformation. The solution of over-segmentation problem in the traditional watershedalgorithm, watershed’s seed region and the line of the watershed area are respectivelypre-calibrated. And then the watershed algorithm is re-used to obtain better results.Finally, the author tests the size distribution system, and analyses three differentrock images’ statistical results. Simulations show that the accuracy reaches to morethan90%by the proposed watershed segmentation method for particle size analysis.Meanwhile, there are some improvements deficiencies in the present system. Thesuccessful development of the system will actively promote the application of imageprocessing technology in the blasting mine rock production.
Keywords/Search Tags:Muck piles, Image processing, Watershed transformation, Sizedistribution
PDF Full Text Request
Related items