摘要
在电子经纬仪测量过程中,点阵图像经常出现不规则的形状,导致识别误码率很高.基于BP神经网络技术,提出了一种适合各种形态点阵特性的识别方法,以增强对不规则点阵和图像噪声的适应能力,从而提高点阵识别的精度.实验结果表明,这种方法具有较高的抗干扰性和抗畸变能力,准确率高,满足事后数据处理要求.
Because of irregular shapes of lattice images produced in photoelectricity theodolite, error rate of code is high when it is recognized. A recognition method based on BP neural network is presented, which is adapted to a variety of lattice shapes. It can process irregular lattices and image noise so that a recognition precision of lattice images can be improved greatly. The experimental results show that the method has anti-noise and anti-aberration ability, making lattice recognition rate higher and meeting data post-processing requirement.
出处
《战术导弹技术》
北大核心
2007年第6期53-55,75,共4页
Tactical Missile Technology
关键词
BP神经网络
点阵图像
特征值
识别
BP neural network
lattice image
characteristic value
recognition
