Answers are grounded in this paper and cite their sources.
Rapid, non-destructive monitoring of seedling physiological status is important for smart greenhouse production, but conventional measurements are labor intensive and difficult to integrate with continuous inspection. In this study, an autonomous mobile multispectral sensing system was developed for estimating LYS-4N-referenced SPAD, nitrogen, and moisture indicators in tomato seedling leaves. A LiDAR-based mobile platform with hardware-adaptive close-range echo filtering was used to provide stable motion support for multispectral leaf acquisition. A four-box green consistency criterion was proposed to trigger spectral acquisition only when effective leaf coverage was satisfied, thereby reducing interference from background regions, leaf edges, and local shadows. After white-reference correction, the leaf spectra were processed using Savitzky-Golay smoothing, standard normal variate transformation, detrending, competitive adaptive reweighted sampling, and XGBoost regression. The prediction-set Rp² values for the LYS-4N-referenced SPAD, nitrogen, and moisture indicators reached 0.81, 0.84, and 0.84, respectively, with corresponding residual prediction deviation values of 2.37, 2.54, and 2.54. Field validation showed average absolute differences of 0.55, 0.21, and 0.84 between platform predictions and handheld measurements. These results demonstrate that autonomous mobile multispectral sensing can support rapid physiological assessment of tomato seedlings in greenhouse production.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: