3D Mapping the Nilein Mosque

This Isn’t for Amateurs

To be very honest, trying to sketch a 3D-version of the Nilein Mosque–also referred to as the Nilin Mosque or, literally in English, the Two Niles Mosque–in Omdurman, Sudan was one of the most frustrating and difficult things I’ve done this semester. Before even beginning to create the 3D image, I had to search Google images for photos to use in the matching process, which essentially entails drawing over the image to give it three dimensions. Luckily, during this process, I was able to find an aerial image of the mosque, to help with roof detail. However, one thing I didn’t find until later was images of the interior. Due to the complex structure of the outer faces, the interior seemed critical in providing a more simplified foundation for the rest of the detailing. Of course, I only thought this until I figured out the simplest way to tackle the issue of the mosque’s core would be to create a polygon and pull it up into a 3D-shape.

Here, I encountered yet another desire for more data that I have been unable to find even with the help of university search engines and online archives. Going strictly from pictures, my accuracy in sketching the 3D model has been highly limited. If, however, the building was more famous and blueprints of the building were available, I could more precisely determine, for instance, that this almost circular polygon has 18 sides without having to guess based on the photo’s perspective or exactly how tall the walls are. Being someone who was never gifted at physics or geometry, it was also hard for me to conceptualize the shape of the roof beneath all the detailing and tiling. I initially assumed that the roof was simply a semicircle; this assumption was proved wrong when I made a semicircle on the blue axis (the one that goes up) from one side of the flat top of the polygon to the other side. When compared with the photo, this semicircle was not only significantly higher at the top, but some of the roof tiles were also below the line of the semicircle. Therefore, the roof is either not a perfect semicircle but more of a dome, or the perspective of the photo is skewed. Again, my lack of aptitude in physics and geometry became my downfall.

Indeed, not only when I was trying to map the roof but also when I was trying to map the arched windows and doorways, my inability to recollect angle rules that I learned in high school made the project immensely frustrating and difficult. Rather than depending on solid geometrical knowledge, I engaged in a game of trial and error, which resulted in error much more often than success. Sometime even when I would succeed at creating a satisfactory arc, I would toggle off the photo and realize that the arc had not been mapped on the face of the building, but was floating in its own space instead. Unsure how to resolve this problem, I would usually just delete the arc and start again. 

Amazingly, all of these issues are nothing compared to my difficulty understanding how to tile the roof itself. In fact, the first time I tried tiling the roof, they grotesquely shot off in to space, as you can see below.


On my second try, I continued to face uncertainty with one roof tile appearing as a projection, like a sun visor, from the face of the building instead of slanting upwards and inwards in proper dome-like fashion. To resolve the issue, I basically drew a vertical line in the very center of the polygon, erasing and redrawing until the height matched the photo. Then, I started drawing lines from that center to the edges of the various faces. Finally, I started using rotated rectangles for the roof tiles, which I sometimes got in the right direction and sometimes had to redraw time and time again until I eventually managed to get the rectangle in the right direction. Afterwards, it was necessary to draw a line on the face of the rectangle to cut into the triangular shape of the roof tiles. This process worked relatively well but having a better understanding of how to use the rotation tool effectively would have helped a lot.

To date, the project remains relatively unfinished, for good reason. Since the structure was recently constructed and remains standing, the extra information I need to finish the Nilein Mosque model with a higher level of accuracy is available. Just by visiting the mosque, I could officially determine how many windows or faces are on the mosque. Similarly, either by visiting the mosque or retrieving blueprints, I could use the geometries and measurements to ensure consistency in the model. Indeed, the misalignment of roof tiles when switching between different reference pictures could likely be resolved with any of this other information.

The resultant hot mess of a shape-file is available for your amusement here.

And here’s a quick look: