Fault detection of surveillance camera using fixed-point arithmetic operations

In this paper, we propose fixed-point operations to measure frame quality assessments as features, and then the computed features are analyzed for fault detection of surveillance camera. The proposed operations include fractional addition, fractional multiplication, and fractional division. Two kinds of quality assessments are employed namely structure similarity and average color. The proposed fault detection scheme analyzes the features using rule-based determination, discriminates corrupted frames among normal ones, and determines whether surveillance camera is normal or not. The experiment results will demonstrate that our scheme is efficient in fault detection of surveillance camera based on fixed-point operations.