Determining the Blasted Block Size Distribution Via Radar D G Johnson 1 ABSTRACT Monitoring of the blasted block size distribution (BBSD) is an important part of the mining process. Existing image processing methods for measuring the BBSD are unable to operate fully in areas of adverse environmental conditions. For this reason, alternate methods for measuring the BBSD should be investigated. In this paper, the challenges that arise when employing radar to measure the multi-aspect height distribution of randomly rough surfaces are discussed. A method employing a broadband scatterometer is proposed for determining the BBSD. INTRODUCTION Monitoring of blasted block size distribution (BBSD) is an important part of the mining process. From the BBSD, information about the in situ block size distribution (ISBD) can be inferred and blasting characteristics adjusted, leading to a more uniform BBSD and, hence, a more efficient mining process (Widzyk-Capehart, 2002). The BBSD can be determined with greatest accuracy by manual sieving of a muckpile sample; however, this is a time-consuming and, hence, costly process, and cannot occur in real-time. More recently, low-cost methods of determining the BBSD using image processing methods have been introduced; either utilising still photographs of muckpiles (Latham, 2003), video footage of muckpiles taken from a roving camera, or still photographs taken from a fixed position above a conveyor system (Maerz, 2001) or from a position overlooking the blasted rocks in transport (Maerz, 2004). Each of these methods has their own advantages and disadvantages; however, common to all is the inability of image processing methods to function reliably in harsh environments, where the presence of dust/steam is likely. For many years, radar systems have been used in areas where vision based systems have been unable to operate (Brooker, 2003). It is, therefore, the intention of this body of work to establish whether radar may be employed to generate the BBSD estimates with equivalent or even greater accuracy than vision based methods. Australian Mining Technology Conference 26 - 27 September 2006 133 1. Australian Centre for Field Robotics, The University of Sydney, Sydney NSW 2006. Email: d.johnson@cas.edu.au Imaging of muckpile Imaging of conveyor Imaging in transport Texture variation • • • Perspective/scaling • Lighting variation • • Obscurement by dust/steam • • • Altered BBSD • TABLE 1 Detrimental effects experienced by image processing methods.