Measurement of three-dimensional pavement morphology using binocular reconstruction improved by second-order fusion algorithm

Abstract

Complex illumination interference poses a practical challenge in achieving robust and accurate measurement of three-dimensional (3D) pavement morphology using traditional binocular reconstruction technology. To enhance the measurement effect, the dual constraints of bi-spectral and structured light are introduced. The theoretical model of improved binocular vision algorithm based on second-order fusion is established. A test system with accuracy self-evaluation is also manufactured. The measurement accuracy, anti-light interference ability and measurement effect in real scenarios are further evaluated. The results show that the improved algorithm can achieve the dual functions: 3D morphology measurement and accuracy self-evaluation. By incorporating dual constraints, the improved algorithm enhances both overall and single point measurement accuracy. Meanwhile, the proposed algorithm can maintain measurement stability in the illumination range of 100–1250 Lux and obtain satisfactory test results in real scenarios. This method exhibits potential for effectively detecting 3D morphology in complex scenarios.