A new and practical method to obtain grain size measurements in sandy shores based on digital image acquisition and processing

Abstract

Modern methods for the automated evaluation of sediment size in sandy shores relay on digital image processing algorithms as an alternative to time-consuming traditional sieving methodologies. However, the requirements necessary to guarantee that the considered image processing algorithm has a good grain identification success rate impose the need for dedicated hardware setups to capture the sand surface images. Examples are specially designed camera housings that maintain a constant distance between the camera lens and the sand surface, tripods to fix and maintain the camera angle orthogonal to the sand surface, external illumination systems that guarantee the light level necessary for the image processing algorithms, and special lenses and focusing systems for close proximity image capturing. In some cases, controlled image-capturing conditions can make the fieldwork more laborious which incurs in significant costs for monitoring campaigns considering large areas. To circumvent this problem, it is proposed a new automated image-processing algorithm that identifies sand grains in digital images acquired with a standard digital camera without any extra hardware attached to it. The accuracy and robustness of the proposed algorithm are evaluated in this work by means of a laboratory test on previously controlled grain samples, field tests where 64 samples (spread over a beach stretch of 65km and with grain size ranging from 0.5mm to 1.9mm) were processed by both the proposed method and by sieving and finally by manual point count on all acquired images. The calculated root-mean-square (RMS) error between mean grain sizes obtained from the proposed image processing method and the sieve method (for the 64 samples) was 0.33mm, and for the image processing method versus manual point counts comparison, with the same images, was 0.12mm. The achieved correlation coefficients (r) were 0.91 and 0.96, respectively.