Light is a critical factor regulating plant development and productivity under controlled environment conditions. However, the light conditions are often kept static throughout the cultivation period, potentially overlooking plants' dynamic responses to changing environmental stimuli over time. This study proposes a stage-specific optimization strategy to maximize lettuce growth, based primarily on shoot fresh weight by adjusting red:green:blue (R:G:B) light ratio at different growth stages. After transplanting 2-week-old seedlings, their growth period was divided into an early stage (ES, the first 2 weeks) and a late stage (LS, after 2 weeks). To account for potential carry-over effects, the ES optimization was designed to evaluate how early-stage light conditions influence final growth performance. Response surface methodology was then employed to identify the optimal spectral combinations for each stage. The optimal R:G:B light ratios were determined to be 44.2:55.8:0 for ES and 25.2:57.8:16.9 for LS. These results suggest that excluding B light during ES promotes morphological traits favorable for light interception, presumably at the expense of immediate photosynthetic efficiency, and ultimately supporting enhanced biomass accumulation during LS. A sequential-optimized lighting strategy combining these two stage-specific light ratios was then evaluated against other lighting strategies, including a static-optimized, a reference, two white LED treatments with different color temperatures of 2700 and 5000 K. While the static-optimized treatment with an R:G:B ratio of 77:23:0 produced the highest shoot fresh weight during ES, the sequential-optimized ultimately delivered the greatest biomass by the end of the growth stage. These findings highlight the importance of stage-specific light requirements and demonstrate that dynamic light management aligned with developmental physiology can significantly enhance crop productivity. This study provides a practical framework for implementing adaptive light strategies in controlled environment systems.

Crop growth information is collected through destructive investigation, which inevitably causes discontinuity of the target. Real-time monitoring and estimation of the same target crops can lead to dynamic feedback control, considering immediate crop growth. Images are high-dimensional data containing crop growth and developmental stages and image collection is non-destructive. We propose a non-destructive growth prediction method that uses low-cost RGB images and computer vision. In this study, two methodologies were selected and verified: an image-to-growth model with crop images and a growth simulation model with estimated crop growth. The best models for each case were the vision transformer (ViT) and one-dimensional convolutional neural network (1D ConvNet). For shoot fresh weight, shoot dry weight, and leaf area of lettuce, ViT showed R2 values of 0.89, 0.93, and 0.78, respectively, whereas 1D ConvNet showed 0.96, 0.94, and 0.95, respectively. These accuracies indicated that RGB images and deep neural networks can non-destructively interpret the interaction between crops and the environment. Ultimately, growers can enhance resource use efficiency by adapting real-time monitoring and prediction to feedback environmental controls to yield high-quality crops.