基于像素的模式识别的数据结构

I have an application that constructs random images based on constraints. Different colored pixels are randomly selected and placed in a grid that satisfy all constraints. For example (simplifying), there might be a constraint that says if a blue or green pixel is at (0,-1), and red pixels are at (-1,-1) and (-1, 0), then placing a white pixel is prohibited. These coordinates are vectors from the current placement location (i.e. its neighborhood).I have an application that constructs random im