Abstract:
Concealed weapon detection (CWD) is a gradually more important topic in the general area of law enforcement and it appears to be a critical technology for dealing with terrorism, which appears to be the most considerable law enforcement problem for the next decade. This syndicate will endeavor a method to detect concealed weapon by using digital image processing techniques. The project is divided in to two parts i.e. image processing and video processing. Image processing is based on 5 keys steps that will take original RGB images and IR (infrared) images as an input which will be preprocessed. After preprocessing features are extracted to get the binary images of concealed weapons and their edges are detected to get the shape of weapons. For video processing first IR video will make converted in to frames and discrete wavelet transform is applied on frames and finally sobel filter for edge detection and homomorphic filter for increasing contrast of a canceled weapon. Experimental results will be evaluated on bases of two parameters i.e. visual inspection and computational cost.
