A multi-visual measurement method for ultimate tensile elongation of steel strand based on spatial characteristics

Abstract

Prestressed steel strands exhibit remarkable strength and relaxation properties, making them essential in engineering projects such as bridges and high-rise buildings. Traditional contact-based measurements using strain gauges typically require adhesive bonding, which may pose mechanical injury risks to operators and surface damage to the strands, and their accuracy is generally limited to ±0.05 mm. To address these limitations, this paper proposes a novel multi-camera machine vision -based measurement technique that utilizes a synchronized four-camera stereo vision system and a data-driven spatial deformation correction model to measure the elongation rate of prestressed steel strands under ultimate tensile loads. Unlike existing contact or single-camera systems, our approach captures real-time transverse and longitudinal deformations through synchronized multi-view imaging and processes them via an advanced spatial deformation model. Uniaxial tension tests were conducted on nine repeated samples of the same steel strand material, divided into three separate batches. The repeated tests were performed under consistent conditions to ensure reproducibility and reduce experimental randomness. All specimens shared the same diameter, twisting structure, and were tested within the ambient temperature range of 15 °C to 35 °C.Compared with reference strain gauge data (baseline error ±0.05 mm), the proposed method achieved a deviation of less than 0.02 mm across all scenarios. Its non-contact, high-precision measurement capability and adaptability to multiple strand types demonstrate feasibility for both laboratory testing and in-field inspections.