Realization of Askant Glance Camera Vision System by Using Extended Hough Transform

We propose a new vision mechanism which can saccadicly move its fixation point for detecting straight lines within tlie image frame. Tliis system, askant glance camera vision ( AGCV ) system, can extract straight lines precisely around the fixation point and ratlier coarsely at tlie periphery of the fixation point. Tliis camera system works in such a way that tlie fixation point catches tlie moving object by tlie information of the center of b~avity of edges, and the edges of tlie moving object can be extracted ratlier finely than the edges of the sleeping object. Tliis system demonstrates a model of tlie center fine / peripheral coarse characteristics of human vision. This system is constn~cted by a sub-system which uses extended Hougli transform filnctions and another sub-system which tracks an object and shifts its fixation point to tlie center of gravity. In tliis paper, we show some experimental results to sliow tlie performance of AGCV system. basic configuration of AGCV system and how to track the moving objects in a manner of askant glance. Finally we present some experimental results to demonstrate the effectivity of the AGCV system. Although this camera vision imitates sidelong glance, it is different from such a camera vision that pursues an object by moving its camera head. AGCV system extracts straight lines precisely around the fixation point and rather coarsely at tlie periphery of tlie fixation point in a software contrivance using an extended Hougli transform ( EHT ) fiinction. In section 3, we sliow the basis theory of EHT fiinction, and in section 4, we show an example of AGCV system which tracks moving objects. In section 5, the concept of the askant glance is extended to the angular sensitivity for line detection. 2 The Concept of AGCV System We wish to compose tlie following AGCV system. AGCV system is one of tlie center fine / peripheral coarse camera vision systems that provides more