摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑每日新闻-机器人方面的最新数据在一份新的报告中呈现。根据NewsRx记者来自胡恩加里莫森马加罗瓦尔的新闻报道，研究表明，“在农业领域，随着精准农业越来越多地使用机器人来监测作物，除草和HAR归属机器人的使用扩大了对计算机视觉的需求。”新闻记者从四川伊斯特万大学的研究中得到一句话:“目前,大多数研究人员和公司都使用基于CNN的深度学习来解决这些计算机视觉任务。这项技术需要大量由专家标记的植物和杂草图像数据集以及大量的计算资源。”传统的基于特征的计算机视觉方法可以提取有意义的参数,并且只需数据集的十分之一就能获得比较好的分类结果.本文提出了这些方法,并试图确定实现可分辨分类所需的最小训练图像数.我们用5,10,20,40,80,在四类分类系统中，每种杂草160幅图像，分别提取了形状特征、距离变换特征、颜色直方图和纹理特征。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Fresh data on robotics are presented i n a new report. According to news reporting originating from Mosonmagyarovar, Hu ngary, by NewsRx correspondents, research stated, "In agriculture, as precision farming increasingly employs robots to monitor crops, the use of weeding and har vesting robots is expanding the need for computer vision." The news correspondents obtained a quote from the research from Szechenyi Istvan University: "Currently, most researchers and companies address these computer v ision tasks with CNN-based deep learning. This technology requires large dataset s of plant and weed images labeled by experts, as well as substantial computatio nal resources. However, traditional feature-based approaches to computer vision can extract meaningful parameters and achieve comparably good classification res ults with only a tenth of the dataset size. This study presents these methods an d seeks to determine the minimum number of training images required to achieve r eliable classification. We tested the classification results with 5, 10, 20, 40, 80, and 160 images per weed type in a four-class classification system. We extr acted shape features, distance transformation features, color histograms, and te xture features."
NETL
NSTL