Leaf surface roughness measure based on shape from focus

Abstract

Abstract BackgroundSurface roughness has a significant effect on leaf wettability, consequently influencing the efficiency and effectiveness of pesticide spraying application. Therefore, surface roughness measure of plant leaves is conducive to relevant researches. In order to characterize the surface roughness, present methods have to draw support from large apparatus, but they are generally high-cost and not portable enough for field measurement. Methods those instruments even have potentially inherent drawback such as absence of relation between pixel intensity and corresponding height for scanning electron microscope (SEM). ResultsAn imaging system with variable object distance is set up to capture images of plant leaves and a shape from focus (SFF) based method is proposed. These space-variantly blurred images are processed with the proposed algorithm to yield surface roughness of plant leaves. The algorithm mainly improves the current SFF method in image alignment, focus distortion correction, and NaN values introducing to make it applicative for precise 3d-reconstruction and surface roughness measure in small scale. ConclusionCompared with method via optical three-dimensional interference microscope, the proposed method preserves the overall topography of leaf surface and meanwhile achieves superior cost performance. Experiments on standard gauge blocks revealed the RMSE of step was approximately 4.44μm. Furthermore, the focus measure operator SML was supposed to perform best according to Friedman/Nemenyi test.

Code and data availability

The paper reports leaf surface roughness measurements via shape-from-focus imaging of rice, rape, cotton, and tobacco leaves, plus gauge-block calibration data. No public dataset, image, or code repository is provided; the authors state data are available only upon request, and no public URLs are present in the article

No evidence-backed public reproduction asset is currently recorded.