Calculating Naturalness

Celikors and Wells studied which aspects of natural outdoor spaces influence how mentally refreshing it is to be in them, focusing on aspects of environments that can be objectively measured. Their most immediately useful findings relate to how natural a space seemed to viewers.  

The researchers report that when study participants rated images of places (a diverse set of environments were evaluated, from tundra to grassland to rainforest to rural settlements to cities to suburban spaces; buildings shown might be historic or modern) “Green-violet hue (gvHue) describes how similar an image’s average color is to green or violet. . . . Saturation refers to the intensity of the hue and standard deviation of saturation (SDsaturation) is the degree of diversity in an image’s saturation. . . . Straight edge density (SED) was derived by dividing the number of pixels on the straight edges of the image by the total number of pixels in the image. . . . Non- straight edge density (NSED) was derived by dividing the number of pixels on the non-straight edges of the image by the total number of pixels in the image.” Assessments of scene naturalness with significantly and positively correlated with gvHue, saturation, SDsaturation, and NSED, and negatively and significantly associated with SED.

Naturalness was determined by asking participants if the place shown seemed natural and colors with lower saturation appeared more gray-ish. 

Elif Celikors and Nancy Wells.  “Are Low-Level Visual Features of Scenes Associated with Perceived Restorative Qualities?”  Journal of Environmental Psychology, in press, https://doi.org/10.1016/j.jenvp.2022.101800