首页|基于机器视觉的防止电力操作走错间隔图像识别模型

基于机器视觉的防止电力操作走错间隔图像识别模型

扫码查看
构建基于机器视觉的防止电力操作走错间隔图像识别模型,能够有效识别电力操作标志牌,减少因操作人员走错间隔引发的误操作事故.通过机器视觉技术构建防止电力操作走错间隔图像识别模型,基于GoogLeNet结构对防止走错间隔标志牌图像实施图像扫描与特征提取;利用投影法通过标志牌图像的水平分割,获得2个仅有一行字符的子图像,将子图像的单行字符逐个分割,获取多个单个字符;采用基于SIFT变换的目标识别方法对分割后的标志牌图像字符进行目标识别,防止操作人员误入非工作区域.实验结果表明,该模型可清晰准确提取防止走错间隔标志牌的图像特征,可有效避免尺度变换、磨损、污渍等外在因素对图像识别精度及效率的影响.
Image Recognition Model for Preventing Power Operation from Incorrect Intervals Based on Machine Vision
An image recognition model based on machine vision is constructed to prevent power operation from incorrect inter-vals,it can effectively identify power operation signs,and reduce mis-operation accidents caused by operators'incorrect inter-vals.Through machine vision technology,an image recognition model for preventing incorrect intervals in power operation is built,and image scanning and feature extraction are implemented for the sign image of preventing incorrect intervals based on GoogLeNet structure.The projection method is used to obtain two sub images with only one line of characters through the hor-izontal segmentation of the sign image,and then the single character of the sub image is obtained by segmenting the single line of characters one by one.The target recognition method based on SIFT transform is used to recognize the characters in the seg-mented signboard image,so as to prevent operators from entering the non-working area by mistakes.The experimental results show that the model can clearly and accurately extract the image features of the spacing signs to prevent incorrect intervals,and can effectively avoid the impact of external factors such as scale transformation,wear,stains and others,the accuracy and effi-ciency of image recognition can be guaranteed.

machine visionsplit horizonpower operationimage feature extractionincorrect intervaltarget recognition

沈晓兵

展开 >

浙江浙能电力股份有限公司萧山发电厂,浙江,杭州 311251

机器视觉 水平分割 电力操作 图像特征提取 走错间隔 目标识别

浙江能源集团有限公司重大科技项目

ZNKJ-2020-141

2024

微型电脑应用
上海市微型电脑应用学会

微型电脑应用

CSTPCD
影响因子:0.359
ISSN:1007-757X
年,卷(期):2024.40(8)