一种基于神经网络的车身划痕自动检测方法  

An Automatic Detection Method for Body Scratch Based on Neural Network

作　　者：王杨[1] 张平 刘念 郑津[1] Wang Yang;Zhang Ping;Liu Nian;Zheng Jin(School of Computer Science,Southwest Petroleum University,Chengdu Sichuan 610500,China;Network and Information Center,Southwest Petroleum University,Chengdu Sichuan 610500,China)

机构地区：[1]西南石油大学计算机科学学院,四川成都610500 [2]西南石油大学网络与信息化中心,四川成都610500

出　　处：《信息与电脑》2018年第17期23-24,共2页Information & Computer

摘　　要：针对目前人工检测车身划痕所存在的成本高、耗时多、效率低、检测人员主观因素占主导等问题,结合深度学习方法,提出了一种基于神经网络的车身划痕自动检测方案。分析了神经网络的特点及其在车身划痕自动检测中的应用优势,并详细阐述了该方法的处理流程和实现方法。The author proposes an automatic detection scheme of car body scratch based on neural network,combining with deep learning method,in order to solve the problems of high cost,long time,low efficiency and dominant subjective factors of detection personnel.After analyzing the features of neural networks and the application advantages in auto body scratch detection,which elaborates in detail the process and implementation of this method.

关 键 词：划痕 自动检测 深度学习 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]