0

I’m working in a project that I need to do a real-time recognition of text using OCR (in this case I’m trying to use the text recognizer by google) with a DJI drone and an android mobile. So I get the sample of DJI SDK “video stream decoding sample” to understand how the decoding of the data from the drone is converted to a frame that has the output showing on the screen of the application. My problem is that I didn’t understand how the camera of the drone works in the application and how I can get the frames and apply the OCR in real-time and edit it to show the identification.

Renato Souza
  • 13
  • 1
  • You dont have a clear coding question or clear background. Assume you work on commercial drone, so get Windows SDK then integrate the windows AZURE OCR with it. Convert the image and send to online query. I did a similar project before for a DJI and Microsoft collaboration show in Singapore. Takes me 2 weeks to get this product done. If you need it to run on the phone, Install MSDK and integrate the Google vision. Convert the image container and send to google server. It is the same concept – Dr Yuan Shenghai Jun 26 '19 at 20:33
  • Found my old video. https://youtu.be/CcndnHkriyA . In my case, it`s all windows header. in your case you should use NSMutableData *downloadedImageData or NSMutableData data. Restructure the data and upload to google vision OCR https://cloud.google.com/vision/docs/handwriting. cast the NSMutableData(DJI) into ByteString imgBytes(google) then follow the example – Dr Yuan Shenghai Jun 27 '19 at 11:14

0 Answers0