| Related citation: | Bowei Xu,Cungui Yu,Jianlin Zhong,Qing He.Real-Time Feature Detection for Booster Recovery:An Adaptive Harris Framework Integrating Scharr Vectorized Gradients and Dilation-Based NMS[J].Journal of Harbin Institute Of Technology(New Series),2026,33(3):15-31.DOI:10.11916/j.issn.1005-9113.25027. |
|
| Author Name | Affiliation | | Bowei Xu | School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094,China | | Cungui Yu | School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094,China | | Jianlin Zhong | School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094,China | | Qing He | Shanghai Aerospace Systems Engineering Research Institute, Shanghai 201109, China |
|
| Abstract: |
| In rocket booster recovery missions, robust visual perception is critical for real-time localization and landing control, but complex environments, including noise, illumination variations, and occlusion, severely degrade the accuracy, repeatability, and localization precision of traditional corner detectors, often resulting in false detections, missed keypoints, and reduced real-time reliability.To address these challenges, an Adaptive Robust Harris (ARH) corner detection algorithm is proposed in this study. The method integrates four innovations: 1) Scharr operator-based gradient computation for enhanced edge response; 2) An adaptive thresholding via statistical analysis of Harris responses to improve noise robustness; 3) A hybrid approach combining dilation-accelerated Non-Maximum Suppression (NMS) and sub-pixel refinement using OpenCV's goodFeaturesToTrack, reducing redundancy and achieving precise localization (average displacement: 0.0458 pixels); 4) Vectorized computation replacing explicit loops, optimizing runtime by 40%. Experimental validation demonstrates ARH's superior performance: high repeatability (≈0.85) under Gaussian noise, 20% feature reduction in motion blur (vs. 21% for Harris and 82% for SIFT), and stable keypoints under rotation. Computational complexity analysis reveals that adaptive thresholding is optimized from O(N) to O(1), while NMS efficiency improves by 50%. These advancements position ARH as a high-precision, real-time solution for vision-critical tasks in aerospace applications, such as tracking rocket boosters and providing landing assistance. |
| Key words: Adaptive Robust Harris (ARH) Scharr operator adaptive thresholding Non-Maximum Suppression (NMS) complex environments rocket booster recovery |
| DOI：10.11916/j.issn.1005-9113.25027 |
| Clc Number:V448,TP391 |
| Fund: |
