Scattering Characteristics of 905 nm Laser on Asphalt Pavement under Water Coverage
With the continuous development and advancement of intelligent transportation systems,autonomous driving and assisted driving technologies are increasingly maturing.Lidar has emerged as one of the key sensor components in this domain.Lidar is an optical sensor that detects target information from scattered light,where the scattering characteristics of the target significantly influence its detection performance.In the realm of smart driving,vehicles can extract surface features of the pavement from the scattering characteristics of the environment,thus enabling the detection and analysis of road conditions.A critical aspect of studying pavement scattering characteristics involves developing an accurate and efficient computational model for the scattering process.Early research by foreign scholars has yielded a series of classic methods for rough surface scattering processes.More recently,some domestic scholars have introduced partial scattering models to enhance the calculations of pavement scattering characteristics.However,traditional computational methods are typically applicable to radar bands in remote sensing scenarios and come with stringent application conditions.Some numerical methods struggle with computational efficiency and accuracy when handling large-scale targets.To address these challenges,this paper introduces a geometric optics ray tracing method that integrates analytical approximations with numerical methods to model the laser scattering characteristics of water-covered pavements.This hybrid approach allows for the precise simulation of laser scattering characteristics before and after the pavement is covered with water,which can be utilized for feature extraction and road type perception.This innovative method not only enhances the modeling accuracy of laser scattering characteristics but also improves computational efficiency,making it adaptable to various application scenarios.Specifically,this paper investigates the scattering characteristics of asphalt pavement at a 905 nm wavelength when covered with water.Initially,a microstructure model of the rough asphalt surface is developed using linear filtering techniques.Subsequently,a geometric optics ray tracing method is employed to establish a computational model for light scattering from rough asphalt surfaces.Building upon this,the model is refined to account for occlusion and multiple scattering effects during the scattering process,thereby enhancing its applicability across various levels of surface roughness.Further,the model is utilized to calculate the laser scattering intensity distribution for asphalt surfaces both prior to and following water coverage.The accuracy of the model is rigorously validated through laser scattering intensity measurement experiments.Finally,the study explores the variations in scattering characteristics under different incident angles as a function of surface roughness and water cover depth.The research findings indicate that the results obtained from the model established in this study closely align with the experimental measurement results,demonstrating that the method can compute the laser scattering characteristics of asphalt pavements effectively.For dry asphalt pavements,when the incident light is at a small angle of incidence,both s-polarized(s-)and p-polarized(p-)light exhibit peaks near the specular direction;however,the peak of the s-light shifts slightly towards larger angles,while the peak of the p-light shifts towards smaller angles.As the angle of incidence increases,the s-light scattering peak moves towards the larger angle specular direction,with the leading edge of the peak becoming steeper.Meanwhile,the p-light begins to exhibit two distinct peaks,with the Brewster angle appearing at approximately 46°.When the asphalt pavement is covered with water,the water layer significantly attenuates the scattered light,reducing the scattering intensity to about one-fifth of its pre-water state.Under these conditions,the scattering intensity distribution curve of the s-light shifts to the right(towards larger angles)with increasing incident angle,but the rise rate of the leading edge of the peak is slower compared to the dry condition.For the p-light,as the angle of incidence increases,the Brewster angle disappears in the scattering intensity distribution curve.Overall,this paper focuses on common water-covered asphalt pavements and utilizes a geometric optics ray tracing method to construct a laser scattering model for such conditions.The accuracy of this model is validated through experiments.Based on this model,the study employs numerical computations to analyze the influence of factors such as surface roughness and incident light polarization state on the scattering intensity at different incident angles.The results indicate that the distribution curve of the scattering intensity gradually approaches a cosine distribution as the surface roughness increases.Additionally,when the root mean square slope exceeds 0.5,an enhanced backscattering trend emerges.There are significant differences in the scattering characteristics of asphalt pavements before and after they are covered with water.After being covered with water,the width of the scattering intensity distribution curve increases,the Brewster angle of p-light disappears,and the difference between the peaks of s-light and p-light diminishes.Furthermore,covering the asphalt with water leads to a sharp peak in scattering intensity in the mirror reflection direction of the incident light,while the intensity of the peaks outside this sharp peak decreases significantly,amounting to approximately one-fifth of the original peak value.This difference increases with the increase in the incident angle.These research findings are of significant importance for the development of environmental perception technologies in intelligent driving.They provide theoretical and experimental data support for pavement condition perception techniques and the optimization of vehicle-mounted optical sensors.
NETL
NSTL
万方数据
