3D Reconstruction of Maize Leaves Based on Virtual Binocular Visual Technology

A method of three-dimensional reconstruction of maize leaves based on binocular stereovision was proposed. This method uses two parallel and relatively fixed cameras to form a binocular stereo vision system. The improved Zhang Zhengyou plane calibration method is used to calibrate the camera. An effective Harris algorithm is used to collect Harris corner points from the acquired images to extract the characteristics of maize leaves. The feature points are matched by calculating the correlation coefficients and thresholds of the corresponding points in the search window. According to the parallax principle, the three-dimensional point cloud information of maize leaves is calculated, and the preprocessed jade is processed. The point cloud of maize leaves is interpolated by bicubic B-spline curve to reconstruct the three-dimensional mesh surface of the leaves. The simulation results show that the proposed method can effectively and accurately realize the three-dimensional reconstruction of maize leaves. The details such as edge undulation and fold of maize leaves can be completely preserved, and the three-dimensional reconstruction accuracy is high.