Application of Heuristic Algorithms in the Tomography Problem for Pre-Mining Anomaly Detection in Coal Seams

Area of Science:

• Medical Imaging
• Geophysics
• Computational Science

Background:

• Computed tomography (CT) often faces challenges with incomplete datasets, particularly for large or inaccessible objects.
• Detecting anomalies in coal seams, such as gas pockets, is critical for miner safety.
• Existing methods represented objects as point sequences, leading to inefficiencies.

Purpose of the Study:

• To present an improved computed tomography approach for incomplete data scenarios.
• To enhance the speed and accuracy of anomaly detection in hazardous environments like coal mines.
• To compare the performance of various heuristic optimization algorithms for solving inverse problems in CT.

Main Methods:

• The study utilizes an improved object representation using sets instead of sequences to eliminate duplicates.
• It addresses the inverse problem by minimizing an objective function.
• Five heuristic algorithms were evaluated: Aquila Optimizer (AQ), Firefly Algorithm (FA), Whale Optimization Algorithm (WOA), Butterfly Optimization Algorithm (BOA), and Dynamic Butterfly Optimization Algorithm (DBOA).

Main Results:

• The revised representation using sets of objects significantly speeds up the reconstruction process.
• The Dynamic Butterfly Optimization Algorithm (DBOA) demonstrated superior performance compared to other tested algorithms.
• The developed approach effectively detects anomalies, such as gas tanks, in incomplete CT datasets.

Conclusions:

• The proposed computed tomography method offers a faster and more efficient solution for incomplete data problems.
• DBOA is identified as the optimal heuristic algorithm for this specific inverse problem.
• This research contributes to improved safety in mining through enhanced anomaly detection capabilities.