远洋捕捞渔船电子监控视频文字信息提取

计算机图像处理技术发展迅速,基于OCR(Optical Character Recognition)的图像文字提取得到广泛应用。在远洋渔业捕捞中,EMS(Electronic Monitoring System)逐渐开始使用,视频中文字信息提取是系统的重要功能之一。通过左右舷、前甲板、后甲板安装的4个摄像头,获取渔船作业的视频影像,再利用jTessBoxEditor软件生成图片样本,人工添加文字标签,然后通过tesseract软件训练标注的样本,制作traineddata字库,最后用Python编程提取远洋捕捞渔船监控视频中的时间、位置、航速、航向等信息,提取的960条记录,经与原图片对比全部正确,提取的文字信息可为视频存档和渔业管理提供参考。

Text extraction from electronic monitoring video of ocean fishing vessels

In recent years,computer image processing technology has developed rapidly.OCR(Optical Character Recognition)based image text extraction is widely used.EMS(Electronic Monitoring System)has been gradually used in the fishing of ocean fisheries.The video images of fishing boat operation can be obtained through four cameras installed on port,starboard,front deck and rear deck respectively.Text information extraction in video is one of the main functions of the system.In this paper,jtessboxeditor is used to generate image samples.Text position and border are labeled manually.Then,the labeled samples are trained with Tesseract software.Then the output trained data library file can be got.Finally,program is written in Python to call library file.The time,position,speed,heading and other information are extracted from the monitoring video of ocean fishing vessels.The 960 records extracted are all correct compared with the original pictures.The extracted text information can provide reference for video archiving and fishery management.