Purpose: Measuring continuous root growth with less time- and resource-intensive methods is challenging for plant ecologists. We propose an approach where photos of fine root growth were taken with smartphones and analyzed by the deep learning method-based program RootPainter. Methods: Picea abies saplings were grown in transparent boxes in growth chambers at either moderate (mRH) or elevated air humidity (eRH) and in nitrate (NO₃⁻) or ammonium (NH₄⁺) dominated soil. To evaluate the partitioning of roots, we analyzed both fine root projection area (PA) of total and young white roots. We aimed to measure the ability of RootPainter to pick up on variation in fine root growth caused by changes in air humidity and soil nitrogen source. Results: Trees growing at mRH-NH₄⁺ had the highest total PA, 9.4 ± 1.9 cm², while the lowest was in trees growing at eRH-NO₃⁻, 3.9 ± 0.6 cm². The young root PA was highest at the beginning of the experiment and was affected by the initial soil nitrogen source. The older (brown) root PA increased over time. At mRH, the relative growth peak of root browning followed the peak of white roots, while at eRH, the peaks coincided. We found strong agreement between the automatic segmentation and manual annotation (F-measure = 0.88); the error measures showed no treatment-specific bias. Conclusions: We conclude that the increased air humidity reduced the fine root growth and diminished sequential developmental peaks. The combination of smartphone images and RootPainter gives reliable results and is easy to use in future plant growth manipulation experiments.