Objectives Root is an important plant organ of obtaining nutrient from soil. Therefore, how to obtain root parameter values nondestructively and efficiently is a current research hotspot. With the rapid development of high-throughput imaging technology, nondestructive measurement based on multi-view images has become a new method to study root system architecture. In this study, precision of the root multi-view imaging system combined with GIARoot parameter analysis platform was accurately assessed. Then, effects of different N levels on root system architecture of tomato seedlings were analyzed dynamically and quantitatively. This multi-view imaging system can provide the basis for further study on the interaction of root system architecture and mineral elements.

Methods Three N treatments were designed in the experiment, namely, N concentrations of 4, 12 and 20 mmol/L, indicating as N4, N12 and N20. Under the hydroponic condition, “Zhong Za 109” tomato seedlings were planted for 16 days in transparent glass column. The self-designed root multi-view imaging system was used to obtain daily image sequences around the growing root system. Then, root characteristic parameters along the plant growth were quantitative calculated with the GIARoot platform according to the photographed root system nondestructively. At the final measurement (16 d after the planting), destructive measurements were done on root system and analyzed with WinRHIZO Pro software, and the results were compared with GIARoot plantform.

Results There were no significant differences on each root characteristic parameter between the GIARoot and WinRHIZO Pro ( P > 0.05). The slopes of the linear regression between the parameters using the two methods were between 0.96 and 0.99, and values of R² were all 0.99 and REs ranged from 2.95% to 12.69%. The RMSEs of network length, network surface area, network volume and average root diameter were 44.73 cm, 4.96 cm², 0.09 cm³ and 0.05 mm, respectively. Each root characteristic parameter of the N12 was the largest on the16th day among the treatments of N4, N12 and N20. The network length, network surface area, network area and network volume of the N20 were 14.2%, 13.2%, 35.8% and 27.7%, respectively, higher than those of the N4. However, the maximum number of roots and the number of primary lateral roots of the N4 were 28.2% and 30.4% higher than the N20, respectively. In addition, significant differences were found on the network length, network surface area and network area between the three N treatments on the 4th day ( P < 0.05), and the values of them of the N12 were 113.9%, 153.7% and 113.8% higher than the N20, respectively. Meanwhile, significant differences were found on the network volume and the maximum number of roots between the three N treatments on the 12th day ( P < 0.05), and the values of the N12 were 57.0% and 117.9% higher than the N20, respectively. However, there were no significant differences on the average root diameter on the 16th day ( P > 0.05) and all of them were between 0.42 and 0.54 mm.

Conclusions The multi-view images method combining the root multi-view imaging system with GIARoot platform, can nondestructively obtain the root characteristic parameters. The N effects on the root architecture characteristics showed that increasing N concentration within certain range could promote root growth of tomato seedlings. However, the N concentration of 20 mmol/L had an inhibition on tomato seedling root growth, particularly on the number of lateral roots, and N concentration had the least influence on the growth of root diameter.