Greenhouse crop digital-imaging method based on ARM and binocular vision

Abstract

The invention relates to a greenhouse crop digital-imaging method based on ARM and binocular vision. The greenhouse crop digital-imaging method includes steps of S1, calibrating a binocular video camera via a checkerboard calibration method, acquiring left-right images of greenhouse crops by the binocular video camera; S2, correcting the left-right images of the greenhouse crops of the step S1 stereoscopically by the ARM embedded platform and then matching by the improved self-adaptive weight stereoscopic matching algorithm to obtain an initial disparity map, and optimizing disparity of the initial disparity image to obtain a complete disparity map; S3, acquiring a three-dimensional point cloud of the greenhouse crops by means of the triangle ranging theory by an upper computer according to the complete disparity map of the step S2, and rendering the three-dimensional point cloud according to the left-right images of the greenhouse crops of the step S2 and displaying and storing. Compared with the prior art, the greenhouse crop digital-imaging method has the advantages that real-time three-dimensional point cloud information of the greenhouse crops can be obtained, so as to reflect growth conditions and provide control reference based on vision for automatically controlling the greenhouse environment, data accuracy is high, cost is low and the like.

Description

technical field [0001] The invention relates to the technical field of digital imaging, in particular to a digital imaging method for greenhouse crops based on ARM and binocular vision. Background technique [0002] The purpose of digital imaging of greenhouse crops is to acquire and measure the morphology and characteristics of crop plants (or organs) in digital images and three-dimensional point clouds, and analyze and process them. Its significance is that it not only facilitates the analysis of extremely complex and long-term life science research in a quantitative space-time coordinate system, shortens the research cycle, but also directly obtains quantitative crop growth observation results. The research on digital imaging method of greenhouse crops can provide visual basis for modern greenhouse environment control system, and can also provide reliable information source for robot picking system. [0003] At present, the digital research methods of crops are divided i...

AI Technical Summary

Problems solved by technology

Although the two-dimensional digital imaging and image processing technology is low in cost and easy to implement, the algorithm design has limitations, and it loses the morphological depth information of the plant and lacks authenticity.

This algorithm improves the accuracy of disparity by assigning reasonable weight values ​​to the pixels of the matching window by using color similarity and spatial proximity, but it is difficult to solve pixel matching in low-texture and repeated-texture regions.

Images