基于人工智能的无人机测绘遥感图像信息提取方法  被引量:6

Image information extraction method of UAV mapping and remote sensing based on artificial intelligence

在线阅读下载全文

作  者:陈立娜 李真 宋辉 CHEN Lina;LI Zhen;SONG Hui(Neixiang County Power Supply Company of State Grid Henan Electric Power Company,Neixiang 474350,China)

机构地区:[1]国网河南省电力公司内乡县供电公司,河南内乡474350

出  处:《电子设计工程》2023年第24期181-185,共5页Electronic Design Engineering

摘  要:针对遥感图像信息提取过程中,因训练样本过大而导致提取结果不精准的问题,提出了基于人工智能的无人机测绘遥感图像信息提取方法。根据每一张图像数据的归一化指数构建图像信息提取模型;采用人工智能的机器学习卷积过程对图像信息进行降维处理;融合图像特征,获取概率特征图,构建目标相对优属度矩阵,共享卷积过程中的权值,实时更新机器学习的判别参数。使机器学习过程与判别过程平衡,将卷积得到的特征图连接起来作为判别依据,判别图像真假,由此提取图像真实信息。引入一个模型复杂性惩罚项,控制训练样本数量,实现无人机测绘遥感图像信息提取。实验结果表明,所提方法提取精度最高为0.93,损失程度最高为0.22,该方法信息提取精准度较高。Aiming at the problem of inaccurate extraction results caused by too large training samples in the process of remote sensing image information extraction,a UAV mapping remote sensing image information extraction method based on artificial intelligence is proposed;according to the normalization index of each image data,the image information extraction model is constructed;the machine learning convolution process of artificial intelligence is used to reduce the dimension of image information;the image features are fused,the probability feature map is obtained,the target relative optimality matrix is constructed,the weights in the convolution process are shared,and the discrimination parameters of machine learning are updated in real time.Balance the machine learning process with the discrimination process,connect the convolution feature map as the discrimination basis,judge the authenticity of the image,and extract the real information of the image.A penalty term of model complexity is introduced to control the number of training samples to realize the information extraction of UAV mapping remote sensing image.The experimental results show that the highest extraction accuracy of the proposed method is 0.93 and the highest loss degree is 0.22.This method has high information extraction accuracy.

关 键 词:人工智能 无人机测绘 遥感图像 信息提取 

分 类 号:TN957.52[电子电信—信号与信息处理]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象