This is quite difficult for tesseract because of the complex background/image structure. You have to use other tools for text detection/document layout analysis and then you can try to OCR cropped images with tesseract.
For your case, the better solution will be to use e.g. Microsoft Omnipaser. Their
demo.ipynb produces these results for your image:

parsed_content_list = ['Text Box ID 0: 3',
'Text Box ID 1: heFarmGuy',
'Text Box ID 2: TheFarmGuy',
'Text Box ID 3: catnap far',
'Text Box ID 4: Gladiator fart',
'Text Box ID 5: Gladiator far3',
'Text Box ID 6: TheFarmGuy',
'Text Box ID 7: NEIT_Farm',
'Text Box ID 8: ator far3',
'Text Box ID 9: cat nap farm',
'Text Box ID 10: NEIT_Farm06',
'Text Box ID 11: Q.A farm1',
'Text Box ID 12: arm07',
'Text Box ID 13: (FHn)Alliance Center',
'Icon Box ID 14: the number 3.']