文献类型：期刊文献

中文题名：基于无人机遥感影像的农业种植数量测量

英文题名：Crop Yield Measurement Based on Unmanned Aerial Vehicle Remote Sensing Image

作者：赵芳[1];贺怡[1]

第一作者：赵芳

机构：[1]新乡学院计算机与信息工程学院,河南新乡453000

第一机构：新乡学院计算机与信息工程学院

年份：2021

卷号：51

期号：10

起止页码：1110-1115

中文期刊名：无线电工程

外文期刊名：Radio Engineering

收录：北大核心:【北大核心2020】；

基金：河南省科技厅重点研发与推广专项(科技攻关)项目(212102210405)。

语种：中文

中文关键词：精细农业;农作物产量测量;无人机遥感影像;卷积神经网络;特征金字塔网络;目标检测

外文关键词：precision agriculture;crop yield measurement;unmanned aerial vehicle remote sensing image;convolutional neural networks;feature pyramid networks;target detection

摘要：为了发展精细农业并降低农业种植的人工成本,利用无人机遥感影像实现了农作物产量的智能化测量系统。该系统利用低空无人机拍摄农作物的遥感图像,将图像传入卷积神经网络进行特征提取。将卷积神经网络特征图传入特征金字塔网络提取不同尺度的特征图,再将每个尺度的特征图送入不同形状的滑动窗口处理,将特征图拼接起来。实验结果表明,该测量系统能够准确地测量农作物的数量与位置,测量精度较Faster R-CNN提升了约16.2%。
In order to develop the precision agriculture and reduce the labor cost of agricultural planting,the remote sensing images of unmanned aerial vehicle are used to implement an intelligent measurement system of crop yield.The system utilizes low-altitude unmanned aerial vehicle to capture remote sensing images of crops.Then the captured images are fed into the convolutional neural networks to extract features.Secondly,the feature maps of the convolutional neural networks are delivered to the feature pyramid networks to extract feature maps of different scales,and the feature maps of each scale are transferred to sliding windows with different shapes for processing,finally the feature maps are flatten and concatenated.Experimental results show that the designed system can measure and locate the crops accurately.The measurement precision is improved about 16.2%compared to Faster R-CNN.