摘要
图像重识别技术作为人工智能领域的新技术,近年来已在多个领域得到了广泛应用。本文旨在提出一种基于图像重识别技术的粮仓内粮食数量变化检测方法,利用新兴科技赋能农业生产,实现对粮食数量变化的自动化监测。该方法相比于传统方法拥有受人工干预影响小、部署和维护成本低、灵敏度高等优势,相比于其他机器学习方法,具有精准度高、适应性强的特点。构建了针对仓内粮食图像的图像重识别模型,以此提取不同时间点的仓内粮食图像的高维特征,借助特征相似性判断仓内粮食是否发生了变化。通过实验验证,该方法在粮食数量变化检测方面的精准率达到96.6%,召回率达到96.2%,相比其他基于计算机视觉技术的检测方法有明显提升,为粮仓粮食数量变化的自动化监测提供了一种基于新质生产力的思路和方法,促进农业生产力发展由量变到质变,加快推进农业深度转型升级,实现农业高质量发展。
As a new technology in the field of artificial intelligence,image rerecognition technology has been widely used in many fields in recent years.The purpose of this paper was to propose a detection method of grain quantity change in granaries based on image re-recognition technology,and used emerging technology to enable agricultural production and realized automatic monitoring of grain quantity change.Compared with traditional methods,this method had the advantages of less influence by manual intervention,low deployment and maintenance cost,and high sensitivity.Compared with other machine learning methods,this method had the characteristics of high accuracy and strong adaptability.In this paper,an image re-recognition model for grain images in the bin was constructed to extract high-dimensional features of grain images at dfferent time points,and to judge whether the grain in the bins had changed by using the feature similarity.The experimental results showed that the accuracy rate of this method in detecting grain quantity changes reached 96.6% and the recall rate reached 96.2%,which was significantly improved compared to other detection methods based on computer vision technology,It provided an idea and method based on new quality productivity for automatic monitoring of grain quantity changes in granaries,promoted the development of agricultural productivity from quantitative to qualitative change,accelerating the deep transformation and upgrading of agriculture,and realizing the high-quality development of agriculture.
作者
邵辉
Shao Hui(Inspur Digital Grain Storage Technology Co.,Ltd,250101)
出处
《粮食储藏》
2024年第3期37-42,共6页
Grain Storage
关键词
新质生产力
深度学习
目标重识别
粮库监管图像
仓储
new quality productivity
deep learning
object re-recognition
grain silo supervision images
storage
