A STUDY OF NEAR-INFRARED (NIR) FILTER FOR SURVEILLANCE APPLICATION

Abstract

Lately, most illegal activities occur in the dead of night when most of the surveillance cameras cannot capture movements clearly. Therefore, Near-Infrared (NIR) filter was used to increase the visualization of suspect identification when the image or footage is captured in a dark environment. The objective of this study was to determine the optimum NIR filters based on the stability of the camera calibration parameters and to evaluate the accuracy of mapping. In this study, four NIR filters with different wavelengths (715, 780, 830, and 850 nm) were tested. The investigation comprised: (1) the calibrations of the camera and NIR filters and (2) a case study involving a simulation test for surveillance application. The type of sensor used was a digital video camera (Sony HC5E HDV) and the camera was set up at multiple stations to form a single convergence configuration. The statistical Analysis of Variance (ANOVA) was used in this study to find (differences in) the significance of the NIR imaging in the calibration and three-dimensional (3D) measurement. The results showed that the camera parameters varied for every type of filter used and this influenced the 3D measurement of the object mapping. In summary, the 850 nm NIR filter was the most optimum for surveillance application based on the stability of the camera calibration and the standard deviation in the mapping accuracy.