This code is the coarse to fine image registration method for time-series UAV orthophotos. This registration method is designed to align between time-series UAV orthophotos. If you want to use this code, you have to cite the research paper: "Multi-temporal orthophoto and digital surface model registration produced from UAV imagery over an agricultural field" (doi: https://doi.org/10.1080/10106049.2022.2143913)"

The proposed method aims to minimize 3D geometric error of multi-temporal UAVorthophotos and DSMs without GCPs by referring to the precise 3D coordinates of the reference orthophoto and DSM generated by UAV imagery with GCPs. As shown in Figure 1, the proposed methodology consists of two main steps; 1) coarse tofine image registration and 2) DSM registration. Tie-points between reference and target orthophotos are extracted using the speeded-up robust feature (SURF) algorithm with restricted searching space in the coarse registration step. The mutual information (MI) algorithm is then used in the fine registration step. The mismatched tie-points are removed, and the target orthophoto is registered to the reference one by estimating the affine transformation model. Subsequently, height values of non-vegetation areas are extracted as EIFs using the vegetation index, and outliers are removed through a statistical Z-score method estimated by the difference in height values between the EIFs. Finally, the target DSM is registered to the reference DSM using a linear regression model estimated based on the EIFs.