A COMPUTER VISION APPLICATION IN REAL-TIME TO IDENTIFYING BIG ROCKS WITH APPLICATIONS TO THE MINING INDUSTRY (original) (raw)
Related papers
Computer vision-based rock modelling
This paper deals with the application of computer vision to mining automation. In particular, it presents results from an on-going research project investigating the automation of the rock-breaking process, a process typical of underground hardrock mining. The objective of this research is to investigate the use of computer vision to identify and locate the oversized lumps remaining on a grizzly (metal sieve) after blasting, for the secondary rock-breaking operation. Range images of rocks are acquired in our laboratory using the NRCC/McGill laser rangefinder. Rock-pile images are segmented into parts (rocks) based on their surface characteristics. Finally, superquadric models are fit to the segmented range data to characterize the 3D shape of each rock.
SME Annual Conference, 2019
Recent studies towards dragline excavation efficiency have focused on incrementally achieving automation of the entire excavation cycle. Initial efforts resulted in the development of an automated dragline swing system, which optimizes the swing phase time. However, the system still requires human operation for collision avoidance. For full dragline autonomy, a machine vision system is needed for collision prevention and big rock handling during the \u27swinging\u27 and \u27digging\u27 phases of the excavation operation. Previous attempts in this area focused on collision avoidance vision models which estimated the location of the bucket in space in real-time. However, these previous models use image segmentation methods that are neither scalable nor multi-purpose. In this study, a scalable and multi-purpose vision model has been developed for draglines using Convolutional Neural Networks. This vision system averages 82.6% classification accuracy and 91% detection in collision avoidance. It also achieves an 87.32% detection rate in bucket pose estimation tasks. In addition, it averages 80.9% precision and 91.3% recall performance across terrain recognition and oversized rock detection tasks. With minimal modification, the proposed vision system can be adjusted for other automated excavators
IEE Proceedings - Vision, Image, and Signal Processing, 2005
Size distribution of rock fragments obtained from blasting and crushing in the mining industry has to be monitored for optimal control of a variety of processes before reaching the final grinding, milling and the froth flotation processes. Whenever feasible, mechanical sieving is the routine procedure to determine the cumulative rock weight distribution on conveyor belts or free falling off the end of transfer chutes. This process is tedious and very time consuming, even more so if a complete set of sieving meshes is used. A computer vision technique is proposed based on a series of segmentation, filtering and morphological operations specially designed to determine rock fragment sizes from digital images. The final step uses an area-based approach to estimate rock volumes. This segmentation technique was implemented and results of cumulative rock volume distributions obtained from this approach were compared to the mechanical fragment distributions. The technique yielded rock distribution curves which represents an alternative to the mechanical sieving distributions.
Rock Recognition From MWD Data: A Comparative Study of Boosting, Neural Networks, and Fuzzy Logic
IEEE Geoscience and Remote Sensing Letters, 2010
Measurement-while-drilling (MWD) data recorded from drill rigs can provide a valuable estimation of the type and strength of the rocks being drilled. Typical MWD sensors include bit pressure, rotation pressure, pull-down pressure, pulldown rate and head speed. This paper presents an empirical comparison of the statistical performance, ease of implementation and computational efficiency associated with three machine learning techniques. A recently proposed method, Boosting, is compared with two well-established methods, Neural Networks and Fuzzy Logic, used as benchmarks. MWD data were acquired from blast holes at an iron ore mine in Western Australia. The boreholes intersected a number of rock types including shale, iron ore and banded iron formation. Boosting and neural networks presented the best performance overall. However, from the viewpoint of implementation simplicity and computational load, Boosting outperformed the other two methods.
Image processing applications for customized mining and ore classification
2011
During the mining operation, ore sorting and directing different grade ores to different processing circuits is a manual task in most of working mines, but this work puts a step forward toward automation of this process. The radical development in the area of image and data processing allows speedy processing of the full color digital images for the preferred investigations. In this paper, an approach has been proposed to classify the ores for blast furnace feed, based on the visual texture of the ore particles. The visual texture of ore particles vary with the mineral contents, for example, blue dust, hard ore, soft ore, etc. This information can be quantified by using image processing technique in red, green, and blue color space and first-and second-order statistical analysis. Commonly used Hartlics textural features were calculated along with red, green, and blue color values for 5×5-pixel size windowpanes extracted from five separate images. Results obtained show encouraging accuracy to apply the approach to develop an expert system for online ore quality monitoring to control the ore blending in the feed ore circuits as well as separating gangue minerals present in the feed ores. Matlab 6.5 was used for visual textural analysis and classification.
Mineral Rock Classification Using Convolutional Neural Network
Recent Trends in Intensive Computing
The main aim of the research is to build a model that can effectively predict the type of mineral rocks. Rocks can be predicted by observing it is colour, shape and chemical composition. On-site technicians need to apply different techniques on rock sample in order to predict rock type. Technicians need to apply different techniques on rock samples, so it is a time-consuming process, and sometimes the predictions may be accurate, and sometimes predictions may be false. When predictions are false, it might show a negative impact in several ways for workers and organization as well. We considered an image dataset of rock types, namely Biotite, Bornite, Chrysocolla, Malachite, Muscovite, Pyrite, and Quartz. We applied CNN (Convolutional Neural Network) Algorithm to get a better prediction of different mineral rocks. Nowadays, CNN is mainly used for image classification and image recognition tasks.
Computer vision-based limestone rock-type classification using probabilistic neural network
Geoscience Frontiers, 2014
Proper quality planning of limestone raw materials is an essential job of maintaining desired feed in cement plant. Rock-type identification is an integrated part of quality planning for limestone mine. In this paper, a computer vision-based rock-type classification algorithm is proposed for fast and reliable identification without human intervention. A laboratory scale vision-based model was developed using probabilistic neural network (PNN) where color histogram features are used as input. The color image histogram-based features that include weighted mean, skewness and kurtosis features are extracted for all three color space red, green, and blue. A total nine features are used as input for the PNN classification model. The smoothing parameter for PNN model is selected judicially to develop an optimal or close to the optimum classification model. The developed PPN is validated using the test data set and results reveal that the proposed vision-based model can perform satisfactorily for classifying limestone rocktypes. Overall the error of mis-classification is below 6%. When compared with other three classification algorithms, it is observed that the proposed method performs substantially better than all three classification algorithms.
Proceedings of the 11th International Conference on Signal Processing and Multimedia Applications, 2014
It is described the algorithm for automatic segmentation of colour images of ores, using the methods of cluster analysis. There are some examples illustrated using of the algorithm in the solving of mineral rock recognition problems. Results of studies are demonstrated different colour spaces by k-means clustering. It was supposed the technique of pre-computing the values of the centroids. There is formulas translation metrics colour space HSV. The effectiveness of the proposed method lies in the automatic identification of interest objects on the total image, tuning parameters of the algorithm is a number that indicates the amount allocated to the segments. This paper contains short description of cluster analysis algorithm for the mineral rock recognition in the mining industry. 165 Baklanova O. and Shvets O.. Methods and Algorithms of Cluster Analysis in the Mining Industry-Solution of Tasks for Mineral Rocks Recognition.