Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive Contextual Embedding for Robust Far-View Borehole Detection

Xuesong Liu, Tianyu Hao, Emmett J. Ientilucci

TL;DR

The paper tackles robust detection of densely packed, tiny boreholes in far-view quarry imagery. It introduces an EMA-driven adaptive detection framework with Adaptive Augmentation, Embedding Stabilization, and Contextual Refinement, designed to stabilize features and leverage spatial context in challenging industrial visuals. The experimental results on a proprietary quarry dataset show substantial improvements over baseline YOLO-based detectors, with an all-three-component system achieving up to 74.9% mAP and strong gains across dense, noisy scenarios. The approach offers practical benefits for safety and efficiency in blasting operations and suggests directions for efficient detectors and transformer-based architectures to extend generalization.

Abstract

In controlled blasting operations, accurately detecting densely distributed tiny boreholes from far-view imagery is critical for operational safety and efficiency. However, existing detection methods often struggle due to small object scales, highly dense arrangements, and limited distinctive visual features of boreholes. To address these challenges, we propose an adaptive detection approach that builds upon existing architectures (e.g., YOLO) by explicitly leveraging consistent embedding representations derived through exponential moving average (EMA)-based statistical updates. Our method introduces three synergistic components: (1) adaptive augmentation utilizing dynamically updated image statistics to robustly handle illumination and texture variations; (2) embedding stabilization to ensure consistent and reliable feature extraction; and (3) contextual refinement leveraging spatial context for improved detection accuracy. The pervasive use of EMA in our method is particularly advantageous given the limited visual complexity and small scale of boreholes, allowing stable and robust representation learning even under challenging visual conditions. Experiments on a challenging proprietary quarry-site dataset demonstrate substantial improvements over baseline YOLO-based architectures, highlighting our method's effectiveness in realistic and complex industrial scenarios.

Adaptive Contextual Embedding for Robust Far-View Borehole Detection

TL;DR

The paper tackles robust detection of densely packed, tiny boreholes in far-view quarry imagery. It introduces an EMA-driven adaptive detection framework with Adaptive Augmentation, Embedding Stabilization, and Contextual Refinement, designed to stabilize features and leverage spatial context in challenging industrial visuals. The experimental results on a proprietary quarry dataset show substantial improvements over baseline YOLO-based detectors, with an all-three-component system achieving up to 74.9% mAP and strong gains across dense, noisy scenarios. The approach offers practical benefits for safety and efficiency in blasting operations and suggests directions for efficient detectors and transformer-based architectures to extend generalization.

Abstract

In controlled blasting operations, accurately detecting densely distributed tiny boreholes from far-view imagery is critical for operational safety and efficiency. However, existing detection methods often struggle due to small object scales, highly dense arrangements, and limited distinctive visual features of boreholes. To address these challenges, we propose an adaptive detection approach that builds upon existing architectures (e.g., YOLO) by explicitly leveraging consistent embedding representations derived through exponential moving average (EMA)-based statistical updates. Our method introduces three synergistic components: (1) adaptive augmentation utilizing dynamically updated image statistics to robustly handle illumination and texture variations; (2) embedding stabilization to ensure consistent and reliable feature extraction; and (3) contextual refinement leveraging spatial context for improved detection accuracy. The pervasive use of EMA in our method is particularly advantageous given the limited visual complexity and small scale of boreholes, allowing stable and robust representation learning even under challenging visual conditions. Experiments on a challenging proprietary quarry-site dataset demonstrate substantial improvements over baseline YOLO-based architectures, highlighting our method's effectiveness in realistic and complex industrial scenarios.
Paper Structure (16 sections, 13 equations, 2 figures, 6 tables)

This paper contains 16 sections, 13 equations, 2 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Method
  3. Adaptive Augmentation
  4. Embedding Stabilization
  5. Contextual Refinement
  6. Overall Objective
  7. Experiments
  8. Dataset and Experimental Setup
  9. Effectiveness of Proposed Components
  10. Conclusion
  11. Component-wise Comparisons
  12. Adaptive Augmentation (AA)
  13. Embedding Stabilization (ES)
  14. Contextual Refinement (CR)
  15. Model Size Sensitivity Analysis
  16. ...and 1 more sections

Figures (2)

  • Figure 1: Overview of our adaptive detection method illustrating integration points of Adaptive Augmentation (Sec. 2.1), Embedding Stabilization (Sec. 2.2), and Contextual Refinement (Sec. 2.3).
  • Figure 2: Qualitative comparison of detection results: Ground Truth, Baseline YOLO (Base), and incremental addition of proposed components. Our complete model (AA + ES + CR) shows robust and accurate detections, especially in challenging conditions.