摘要
本文以某项目的光电探测装置为硬件载体,以某次目标特性测试试验取得的图片为样本,介绍了浅层神经网络和卷积神经网络在目标图像的预处理与图像特征提取过程中的作用。对两种算法的正确识别概率进行了比对,证明了卷积神经网络能有效提升目标识别概率。
In this paper,the photoelectric detection device of a certain project is used as the hardware carrier,and the pictures obtained in a certain target characteristic test experiment are used as samples to introduce the role of shallow neural networks and convolutional neural networks in the process of target image preprocessing and image feature extraction.And compares the correct recognition probabilities of the two algorithms.
作者
王涛
吕鑫
Wang Tao;Lv Xin(Beijing Mechanical Equipment Research Institute,Beijing 100854,China)
出处
《科学技术创新》
2022年第7期17-20,共4页
Scientific and Technological Innovation
关键词
目标识别算法
浅层神经网络
卷积神经网络
Target recognition algorithm
Shallow neural network
Convolutional neural network
