Sediment imaging and factorized Haar wavelet transformation techniques for characterizing particle size distributions of granular soils

Abstract

The particle size distribution (PSD) significantly influences various mechanical behaviors of granular soils. Advances in automation and camera technology allow for a rapid, green, and accurate PSD determination. This study develops a new sediment imaging system, mainly consisting of a sediment column filled with water and a line-scan camera driven by programmable logic controller, which can automatically capture high-resolution soil segregation images. The soil particles settle through the sediment column and segregate by size, which is scanned to generate sediment images. Naturally, the coarsest particles settle at the bottom of the image, while the finest particles are located at the top. Therefore, the analysis of the sediment image enables the determination of PSD. The higher sediment column, despite of poor operationality, yields better particle segregation quality and thus the more accurate PSD result. This study proposed a factorized Haar wavelet transformation (FHWT) algorithm that can account for less perfect segregation of particles by size, enabling to shorten the height of sediment column. The proposed hardware and FHWT are evaluated by six natural sands with different colors, internal textures, and geological origins. An excellent match between image-based PSDs and sieving-based PSDs are observed. Based on FHWT, four sediment columns with different heights of 2.1 m, 1.5 m, 1.0 m, 0.5 m are tested. The results indicate that columns higher than 1.0 m yield acceptable particle segregation results. Furthermore, the proposed FHWT is also adaptable to process high-resolution images captured by advanced cameras.