如何从命令行使用 tesseract 在打开的窗口内读取内容？
我正在尝试从命令行使用 tesseract 对打开的窗口的内容运行 OCR。特别是，我愿意阅读当前打开的记事本窗口中输入的文本。
我在这里阅读了文档和维基：http://code.google。 com/p/tesseract-ocr/w/list
但我没有找到任何对我这个项目有帮助的东西，此外我也在这里搜索了类似的问题（有很多关于 OCR 的问题），但似乎什么也没有工作/适用于我的情况。
可行吗？
我主要是一名 PHP 编码员（编码只是为了好玩），没有非网络语言的经验。
提前致谢。
I'm trying to use tesseract from command line to run OCR on the content of an opened window. In particular I'm willing to read the text typed into a current opened Notepad window.
I've read the documentation and the wiki here: http://code.google.com/p/tesseract-ocr/w/list
but I didn't find anything that helped me in this project, further more I've also searched here for similar questions ( there are many about OCR) but nothing seems to work/ be applicable in my case.
Is it feasible?
I'm mainly a PHP coder (coding just for fun) and have no experience in non-web languages.
Thanks in advance.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
Tesseract 旨在将 TIFF 图像作为输入，并且对 Windows 或屏幕设备上下文一无所知。因此，您需要添加代码来定位记事本窗口的窗口句柄，根据 Windows 报告的当前窗口大小执行屏幕捕获和剪辑窗口，并将生成的图像保存到文件中。该图像很可能是黑白的，这将使 OCR 变得更容易，因为我怀疑 Tesseract 2.0 只适用于黑白图像。下一个问题是 Tesseract 由于源图像的 DPI（分辨率）较低而导致结果不佳。
为了评估您的方法的适用性，我将通过打开记事本、截取屏幕截图、在 MSPaint 中打开屏幕截图、剪辑您想要 OCR 的文本、将剪辑的图像保存为 TIFF 或 BMP 并将此文件发送到 Tesseract 来执行一些手动测试。如果结果不如您需要或期望的那么好，这可以为您节省大量时间和精力。
Tesseract is designed to take a TIFF image as input and know nothing about the Windows or screen Device Contexts. So you would need to add code to locate the windows handle for the Notepad window , perform a screen capture and clip the window based on the current window size reported by Windows and save the resulting image to a file. This image will most likely be black and white which will make it easier to OCR as I suspect Tesseract 2.0 only works with B/W Images. The next problem will be Tesseract gving poor results due to the low DPI (resolution) of the source image.
To evaluate the suitability of your approach I would perform some manual tests by opening Notepad, taking screenshots, opening the screenshots in MSPaint, clipping the text you want to OCR, save the clipped image to a TIFF or BMP and send this file to Tesseract. This could save you a lot of time and effort if the results are not as good as you need or expect.