It's a damn garbled problem.

After reading JPEG images in liunx, write EXIF information. normally on Windows, only a small portion of Chinese can be correctly written in liunx. for other unrecognized Chinese characters, it will be changed to the word "Duplicate copy". In fact, these three words only correspond to one word for normal Chinese characters. I thought about using outputstream for byte transcoding. however, the format of the generated JPEG file is corrupted due to the loss of bytes. use byte [] to convert char [] to specify the encoding method. no.
Depressed. It has not been done all night. So depressing !!!!!!!!