In this paper, we introduce an imaging technique for capturing 3-imensional smoke density distribution by adopting a three-layer color illumination and a six-layer color illumination for air flow measurements. A camera capable to capture three primary color components (red, green, blue) and a liquid crystal display projector are used as optical equipment. Each color layer is aligned parallel to a photographed plane of the video camera. The present technique assumes two things; three primary color signals are received proportionally to smoke density in each color layer and there is a linear relation between the smoke density and the color signals. These assumptions allow that the smoke density can be obtained from the color signals by solving inverse problem. In the three-layer case, solving inverse problem is possible by these assumptions. In the six-layer color illumination, a color layer illumination consisting of three primary colors and their intermediate colors (yellow, cyan, magenta) are projected toward measurement space. This forms an inverse problem with lack of information to reconstruct the smoke density because the number of independent color signals is lower than the number of color layers. To fix this issue, we apply a method to increase the number of color signals in time direction by switching two different illumination patters. The number of color signals becomes six to be equal to the number of variables to solve algebraically.