Video based dynamic texture recognition under specific illumination conditions

Nowadays dynamic textures recognition is particularly important in different computer vision community tasks in a variety of fields such as urban scenes and forest scenes. The goal of the dynamic textures recognition can be different. Real scenes may include the objects with dynamic behavior because of possible varying illumination, blurring, or weather conditions. Under bad weather conditions the imaging system is degraded to produce low visibility images. Such effects may significantly degrade the performance of outdoor vision systems which relies on image/video. For illumination effects compensation and visual quality enhancement images it is necessary to average pixel intensity increase, expand the range of brightness, image contrast increase and eliminate influence of the additive noise. For the images obtained in adverse lighting conditions imitation in this work Gamma correction, additive Gaussian noise and impulse noise was applied successively. The proposed algorithm employs Multi Scale Retinex with Color Restoration, Laplacian, Gaussian and median filters. For experimental researches of the databases of Bilkent University, Video Smoke Detection, Wildfilmsindia, V-MOTE were used. In addition, the representativeness of the test set is increased by a video sequence, which the authors themselves recorded at night. Experiments on video based smoke detection system based on spatio-temporal local binary pattern were computed. True recognition for smoke in adverse lighting conditions is degraded to 65 %. False rare rejection and false alert errors significantly increase to 34.2 % and 27.2 % respectively. After adverse lighting compensation algorithm work true recognition of smoke regions increases to 94.41 %. This accuracy provides the influence of adverse lighting on a quality of smoke detection is studied. Experimental results show that the proposed method is feasible and effective for video-based dynamic texture analysis in varying illumina- tion conditions