A registration and optimization method for plant point cloud acquisition based on tof camera

Abstract

The invention discloses a plant point cloud acquisition registration and optimization method based on a TOF camera, comprising the following steps: (1) placing the measured plant on a turntable, adjusting the TOF camera angle of view and aligning it with the measured plant; 2) Turn on the turntable to make it rotate automatically, and at the same time, the TOF camera collects the point cloud data of the measured plant at intervals; (3) Perform real-time preprocessing on each frame of point cloud data collected by the TOF camera; Two frames of point cloud data are registered and optimized, and then integrated to obtain complete plant point cloud data; (5) Statistical filtering is used to remove discrete noise in the number of plant point clouds obtained by registration and optimization processing, and the final point cloud data. The method of the invention can effectively solve the stratification phenomenon of local point clouds.

Description

technical field [0001] The invention relates to the technical field of digital agriculture, in particular to a TOF camera-based method for collecting, registering and optimizing plant point clouds. Background technique [0002] At present, a variety of detection methods and equipment have been applied to the three-dimensional structure acquisition of objects in the industrial field, such as based on lidar, based on structured light, based on TOF camera (time of flight camera, Time of Flight), based on multi-eye vision and other principles Scanning equipment has been widely used. [0003] In the field of agriculture, due to factors such as the dynamics and flexibility of plants, and the diffuse reflection properties of plant leaf materials, the research and application of related instruments is relatively late. [0004] Among them, TOF is a simple, easy-to-install, and cost-effective technology. TOF cameras use active light sources to obtain 3D point clouds of plant objects,...

AI Technical Summary

Problems solved by technology

[0009] The above method can improve the registration accuracy and avoid the error caused by the traditional ICP registration algorithm, but the above method has high requirements for the mark detection of the turntable and the accuracy of the rotation axis setting

[0010] In the optimization part, the noise reduction method is usually used to improve the registration results, but this method cannot effectively solve the layering phenomenon of plant leaves, resulting in the obtained plant point cloud data still cannot be applied to subsequent research calculations

Images