Activation light pattern helps detection

In this paper a CNN frameless computing algorithm is considered for feature detection via lighting activation. The uphill region of a convex bump is measured with an infrared active sensor array, when different deficiencies are present. In the case of global lighting (when all of the light sources of the array are shining), due to the applied deficiency (eq. roughness, ditch, or well) the system is unable to settle down to one, common output pattern; however, in the case of an up-moving periodic lighting wave, the system mostly converges to and remains at a specific output pattern. This pattern uniquely identifies the global uphill trend of the observed terrain.