HighlightsA non-destructive, in situ, and low-cost root phenotyping system was developed.The system can collect color images and 3D cloud points of corn roots in soil.When tested in a greenhouse, the scanning process did not cause significant disturbance of corn plants.The results showed significant differences in root growth for different watering treatments and growth stages.Abstract. Plant root phenotyping technologies play an important role in breeding, plant protection, and other plant science research projects. Root phenotyping researchers urgently need technologies that are low-cost, in situ, non-destructive to roots, and suitable for the natural soil environment. Many recently developed root phenotyping methods, such as minirhizotron, X-CT, and MRI scanners, have unique advantages in observing plant roots, but they also have disadvantages and cannot meet all the critical requirements simultaneously. This study focused on the development of a new plant root phenotyping robot, called MISIRoot, that is minimally invasive and works in situ in natural soil. The MISIRoot system mainly consists of an industrial-level robotic arm, a miniature camera with lighting, a plant pot holding platform, and image processing software for root recognition and feature extraction. MISIRoot can acquire high-resolution color images of roots in soil with minimal disturbance to the roots and measure the roots’ three-dimensional (3D) structure with an accuracy of 0.1 mm. In tests, well-watered and drought-stressed groups of corn plants were measured with MISIRoot at the V3, V4, and V5 growth stages. The system successfully acquired RGB color images of the roots and 3D point cloud data containing the locations of the detected roots. The plants measured with MISIRoot and the plants not measured (control) were carefully compared with the results from a hyperspectral imaging facility (reference). No significant differences were found between the two groups of plants at different growth stages. Keywords: 3D point cloud, Low-cost phenotyping, Minimally invasive root measurement, Plant root phenotyping, Robotic arm application, Root imaging.