|
|
|
| Research and Application of Ore Particle Size Image Detection Method Based on Machine Vision Recognition Technology |
| YUAN Long1,4, LIU Zhipeng2, SHI Huadong1, DU Zibin3,4, HAN Chaoyu3
|
1 CITIC Heavy Industries Co., Ltd. , Luoyang Henan 471039, China; 2 China Energy Engineering Group Prefabricated Construction Industrial Development Co., Ltd. , Shenzhen Guangdong 518116, China; 3 National Innovation Intelligent Mining Equipment Research Institute (Luoyang) Co., Ltd. , Luoyang Henan 471039, China; 4 State Key Laboratory of Intelligent Mining Heavy Equipment, Luoyang Henan 471039, China
|
|
|
|
|
Abstract To address the issues of low efficiency and inability to meet the demands of modern mining production associated with traditional ore particle size detection methods, this paper introduces a machine vision recognition-based image analysis methodology for ore size distribution characterization. The acquired images are first denoised using median filtering, which effectively suppresses noise while preserving particle edge sharpness. Subsequently, adaptive threshold segmentation is employed to separate ore particles from the background, and morphological operations are applied to eliminate adhesions and fill internal voids. Building upon this preprocessing, gradient-based segmentation coupled with distance transformation is utilized to delineate boundaries of touching particles. Finally, particle size distribution is quantified by fitting a calibrated “size-volume” model, enabling accurate particle identification and granularity analysis. Experimental validation from a sand and gravel aggregate production line project demonstrates that the ore particle size image detection system built using this method can reliably perform real-time particle size detection. Compared to manual sieving results, the error in the longest dimension of ore particles is controlled within 8%, and the P80 error is controlled within 14%. The detection accuracy meets the requirements for industrial field applications, providing effective data support for evaluating ore processing performance and optimizing production processes.
|
|
Received: 09 November 2025
Published: 19 May 2026
|
|
|
|
|
| No related articles found! |
|
|
|
|
