Characterizing segregation in blast rock piles a deep-learning approach leveraging aerial image analysis

Chengeng Liu,Sihong Liu,Chaomin Shen,Yupeng Gao,Yuxuan Liu

2024-06-06

Abstract:Blasted rock material serves a critical role in various engineering applications, yet the phenomenon of segregation-where particle sizes vary significantly along the gradient of a quarry pile-presents challenges for optimizing quarry material storage and handling. This study introduces an advanced image analysis methodology to characterize such segregation of rock fragments. The accurate delineation of detailed rock fragment size distributions was achieved through the analysis of drone-captured imagery, coupled with the application of an enhanced Unet semantic segmentation model integrated with an expansion-based post-processing technique. The quarry slope was stratified into four vertical sections, with the size distribution of each section quantified via ellipsoid shape approximations. Our results disclose pronounced vertical segregation patterns, with finer particles concentrated in the upper slope regions and coarser particles in the lower. Utilizing relative characteristic diameters, we offered insight into the degree of segregation, thereby illustrating the spatial heterogeneity in fragment size more clearly. The techniques outlined in this study deliver a scalable and accurate method for assessing fragment size distribution, with the potential to better inform resource management and operational decisions in quarry management.

Image and Video Processing,Artificial Intelligence

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the particle segregation phenomenon that occurs during the stockpiling of blasted rock materials, that is, the problem of uneven distribution of rock particles of different particle sizes in the ore pile along the slope direction. This segregation phenomenon poses challenges to the storage and processing of mine materials, especially in optimizing quarry material management and operational decision - making.

Specifically, this research aims to solve these problems in the following ways:

1. **Develop advanced image analysis methods**: Utilize aerial images obtained by unmanned aerial vehicles (UAVs), combined with the enhanced Unet semantic segmentation model in deep learning and the post - processing technique based on region expansion, to accurately depict the size distribution of rock fragments.
2. **Quantify the vertical segregation pattern**: Divide the quarry slope into four vertical regions, and approximate the particle size distribution in each region through ellipsoid - shaped approximation to reveal a significant vertical segregation pattern in which fine particles are concentrated in the upper region and coarse particles are concentrated in the lower region.
3. **Provide a scalable method**: The proposed method can evaluate the debris particle size distribution, providing better information support for resource management and operational decision - making, thereby optimizing the use and transportation of mine materials.

### Formula Summary

- **Calculation formula for equivalent elliptical diameter**:
\[
d = 1.16\sqrt[3]{\frac{4}{\pi}V}
\]
where \(d\) is the equivalent elliptical diameter, and \(a\) and \(b\) are the major and minor axes of the equivalent ellipse respectively.

- **Calculation formula for ellipsoid volume**:
\[
V=\frac{4}{3}\pi ab^{2}
\]

### Method Framework

1. **Overall framework**: From obtaining rock block images to finally counting the particle size distribution, the entire process includes steps such as image acquisition, pre - processing, semantic segmentation, post - processing, and shape estimation.
2. **Image acquisition and processing**: Use UAVs to take high - resolution images and perform resampling and annotation.
3. **Particle segmentation**: Adopt an improved Unet model for semantic segmentation, and fuse the particle body and boundary through the region - expansion post - processing technique.
4. **Segregation feature description**: Calculate the characteristic particle sizes (such as \(d_{10}\), \(d_{50}\), \(d_{90}\)) in each region, and characterize the segregation phenomenon through relative characteristic diameters (such as \(d_{10}/d'_{10}\), \(d_{50}/d'_{50}\), \(d_{90}/d'_{90}\)).

Through these methods, the research provides a scientific basis for the management and transportation of mine materials, which is helpful for improving the quality and efficiency of engineering design and construction.