Surrounding Structure Estimation Using Ambient Light

Abstract

Image measurement technology – widely used in present society – has made substantial progress. It involves processes such as image input, target extraction, and measurement of the extracted region to obtain information from an image. These processes are computationally intensive because they require a large amount of information such as complex features, which is often an obstacle to improving and speeding up image processing tasks. In contrasts, living organisms easily recognize their own surroundings in real time. In cognitive science research studies, for example, visual affordance studies have shown that organisms perceive and recognize their surrounding environment and objects from ambient light, which is formed by reflected and scattered light in the environment. By applying this natural mechanism to image measurement technology, it is possible to obtain the information necessary to recognize the surrounding environment by observing ambient light without necessarily detecting or recognizing the object. In this study, we propose a direct method of assessing the surrounding environment by capturing ambient light as luminance.