Identification of Rock Fragments after Blasting by Using Deep Learning-Based Segment Anything Model

Junjie Zhao,Diyuan Li,Yisong Yu

DOI: https://doi.org/10.3390/min14070654

IF: 2.5

2024-06-26

Minerals

Abstract:Rock fragmentation is an important evaluation indicator for field blasting operations. This paper applies a deep learning-based method, the Segment Anything Model (SAM), to automatically segment rock fragments. To review the SAM's segmentation performance, 83 images of rock fragment collected from the mine site were used as the test dataset. Pixel-level accuracy (PA), intersection over union (IOU), and dice coefficient (Dice) were employed to evaluate the model pixel-level segmentation performance. The results showed that the SAM exhibited excellent segmentation performance on the test data (PA = 94.5%, IOU = 94.4%, Dice = 95.4%). The coefficient of determination (R2) values for the 50% and 80% passing sizes (X50 and X80) were 0.970 and 0.991, respectively, which demonstrated that the SAM could achieve high precision measurement of rock fragmentation. Additionally, the effectiveness of the SAM was further evaluated by comparing it to commercial software, and the generalizability of the SAM was verified on two other datasets. The findings revealed that the SAM not only outperformed the Split-Desktop V 4.0 on the test dataset but also achieved comparable accuracy to previous studies on the two other datasets. The SAM could be regarded as a useful tool to provide fast and accurate feedback for field blasting.

geochemistry & geophysics,mineralogy,mining & mineral processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to use deep - learning methods to automatically, efficiently and accurately identify and measure the size of rock fragments after blasting**.

Specifically, the authors applied an image - segmentation model based on deep learning - Segment Anything Model (SAM) - to automatically segment and measure rock fragments after blasting. Traditional methods for measuring rock fragments usually rely on empirical models or traditional image - processing techniques. These methods have poor generalization ability on different data sets and require a large amount of manually - labeled data. As a pre - trained base model, SAM can achieve zero - shot segmentation without a large amount of labeled data, thus significantly reducing labor costs and improving efficiency.

### Main problems and solutions

1. **Improve the accuracy and efficiency of rock - fragment measurement**:
- Empirical formulas in traditional methods such as the Kuz - Ram model are difficult to generalize to other data sets.
- Although the image - analysis method is simple and efficient, it performs poorly when dealing with images of different brightness, shadows and surface textures.

2. **Reduce the dependence on large - scale labeled data**:
- Deep - learning methods usually require a large amount of labeled data for training, which is both time - consuming and expensive.
- SAM can achieve zero - shot segmentation without the need to prepare training data in advance, reducing the cost and time of data preparation.

3. **Verify the effectiveness and generalization ability of the model**:
- Evaluate the performance of SAM by comparing it with commercial software (such as Split Desktop).
- Use two other data sets to verify the generalization ability of SAM to ensure its stable performance in different scenarios.

### Research significance

By introducing the SAM model, this research not only provides a fast and accurate tool for measuring rock fragments, but also shows the potential application value of deep learning in engineering fields such as geology and mining. The successful application of SAM provides new ideas and technical means for future research in related fields.