The rising global population, in conjunction with the increasing demand, decreasing labor supply, and increasing costs in the agricultural sector, has induced a need for automation in this industry. Many of these tasks are simplified using depth images and are accomplished using the help of 3D sensing technology such as stereo vision and time of flight methods. While there are various merits to these approaches, there is a need for high-speed, high-accuracy 3D profiling approaches in this rapidly advancing industry. Fringe Projection Profilometry is a variation of structured light technology, which has the advantage of having high speed in the kilohertz range, and sub-millimeter accuracy, which could be extremely beneficial for this sector to adopt. In this article, we seek to provide a tutorial on this technology, explain its various principles along with the basic methodology, and expound on its advantages. We demonstrate some example results using soybean roots and spinach leaves to show its utility, discuss potential reasons as to why this has not yet been widely adopted by this industry, review its potential limitations, and examine possible ways those limitations can be addressed so that they do not present a roadblock in its adoption.

This paper proposes a phase-to-depth deep learning model to repair shadow-induced errors for fringe projection profilometry (FPP). The model comprises two hourglass branches that extract information from texture images and phase maps and fuses the information from the two branches by concatenation and weights. The input of the proposed model contains texture images, masks, and unwrapped phase maps, and the ground truth is the depth map from CAD models. A loss function was chosen to consider image details and structural similarity. The training data contain 1200 samples in the verified virtual FPP system. After training, we conduct experiments on the virtual and real-world scanning data, and the results support the model’s effectiveness. The mean absolute error and the root mean squared error are $1.0279$ mm and $1.1898$ mm on the validation dataset. In addition, we analyze the influence of ambient light intensity on the model’s performance. Low ambient light limits the model’s performance as the model cannot extract valid information from the completely dark shadow regions in texture images. The contribution of each branch network is also investigated. Features from the texture-dominant branch are leveraged as guidance to remedy shadow-induced errors. Information from the phase-dominant branch network makes accurate predictions for the whole object. Our model provides a good reference for repairing shadow-induced errors in the FPP system.