For Part 1, I created an arc-shaped wall surface in Grasshopper with sliders controlling the radius, start angle, end angle, and height. The surface was then divided into a grid of nearly square panels using the PanelQuad node, with the U and V counts calculated automatically from the arc length and wall height so the panels stay proportional for a dimensions change. For the image, I used Image Sampler to take a photo and map its colors on the panel grid. The center point of each panel was found using the Area and Surface Closest Point nodes, which returned UV coordinates that were fed directly into the Image Sampler. the wall can be adjusted using sliders for radius, start angle, end angle, and height. The panel grid automatically recalculates to stay proportional with any dimension change. Swapping the input image changes the color pattern across the panels entirely.
For Part 2, I made a sine wall with sliders controlling the length, height, number of waves, and amplitude. A sine wave function was used to make a base curve, which was then extruded vertically to create the S-shaped wall surface. The surface was divided into a grid of panels using the Lunchbox PanelQuad node, with the U and V counts calculated automatically from the curve length and wall height to maintain a consistent proportion across any wall dimension. To drive the panel depths, an Image Sampler was used to read the brightness values of a photo across the panel grid. The same coordinate mapping approach from Part 1 was used to ensure a one-to-one correspondence between the image data and the panels. The brightness values were then remapped to a height range and used to extrude each panel outward along the wall surface normal, with brighter areas of the image producing deeper extrusions and darker areas sitting flatter against the wall.