The ultrascale is definitely pricy for non-industry applications. The FPGA design is probably larger because of the 4x camera pipeline.
Perhaps with a single camera you could port this to fit on a Zynq 7000 footprint with something like a Pynq Z1 or Numato Styx, which are in the $250 hobbies price point for example.
No chance does the Zynq 701x series perform capably for this task: it's weak Artix fabric and there's only ~50kLUT on the common parts. Even fully pipelined designs, especially those that interface with the hard AXI bus, will struggle to clock above 50MHz.
The Zynq Z7030+ chip would probably be able to get the job done but aren't as common - https://www.lcsc.com/product-detail/Programmable-Logic-Devic.... The Kintex 410T-480T are available new <$160 from reputable vendors, used <$50 from less-reputable vendors - they do have the performance (and overall IO) for this task.
Perhaps with a single camera you could port this to fit on a Zynq 7000 footprint with something like a Pynq Z1 or Numato Styx, which are in the $250 hobbies price point for example.