I'm thinking those image urls have same figure like :
What you only need to change the number at ending , e.g, 1.jpg, 2.jpg ..etc. , so you can try to populate those image urls as manually by a query or content transformation .
However, the main cause is in the xpath with 'image_url' element ,it does select 99 elements, so I've revised the xpath to select all , all image urls can be extracted properly.
See the attached new project and all sample data.