基于机器视觉的葵盘面积提取及产量预估试验研究

Experimental Study on the Area Extraction and Output Prediction of Sunflower Plate Based on Machine Vision

产量的获取是实施精准农业的关键,许多研究人员提出了多种向日葵产量监测方法,但测产的精度及适用性还不能完全满足用户要求。为此,提出了采用图像处理的方法对向日葵产量进行监测,并搭建了机器视觉系统的硬件平台对向日葵葵盘图像进行采集,拟定两种图像处理方案,确定最优方法。选用八邻域算法对两种方案标记的最大连通区域进行面积信息计算,通过面积信息与人工计数得到的籽粒数量获取单个籽粒的面积信息,再由多组试验确定单个籽粒面积信息,依据葵盘面积与单个籽粒的面积计算单个葵盘中的籽粒数量,以称取千粒质量的方法获得单盘产量及总产量。

The production of obtaining is the key to the implementation of precision agriculture,many researchers at home and abroad a variety of sunflower production monitoring method is proposed,but made in measuring precision and generality of the model can not fully meet user requirements for intelligent features such as wide application of machine vision technology,therefore,in this paper adopts the method of image processing for sunflower production monitoring in this paper has set up a hardware platform for the machine vision system for sunflower sunflower plate image acquisition,presents two kinds of image processing schemes,determine the optimal method choose eight neighborhood algorithm of two kinds of schemes marking the biggest area information of connected region,from the area of information and artificial counting the grain number for a single grain area in the form of information,again by multiple sets of experiment to determine a single grain area of information,the next kwai plate area with a single grain area computation single kwai plate grain number,finally to weigh and method for a single grain yield and total output of.