.

ISSN 2063-5346
For urgent queries please contact : +918130348310

DETECTION OF MOVING OBJECTS IN BAD WEATHER USING DUAL BACKGROUND ILLUMINATION COMPENSATOR IN COMPARISON WITH FRAME DIFFERENCING, SINGLE GAUSSIAN AND GMM TO MEASURE F-SCORE

Main Article Content

V. Maneesh Reddy , Karthikeyan P R
» doi: 10.31838/ecb/2023.12.sa1.301

Abstract

Aim: By using a background subtraction algorithm the foreground objects are detected. The proposed model performs better comparison algorithms with more f-score and accuracy. The use of this method to detect or have clear images of the objects even in bad weather. Materials and Methods: The total number of 29000 images are checked and subtract the background noises and give clear foreground images like pedestrians, cars, skating . By using algorithms in the matlab application. Result : The proposed algorithm is tested in four video sequences of various illumination conditions. As the datasets containing both gradual and sudden illumination change. The mean f-scores values of Frame differencing, Single Gaussian, Gaussian mixture method and Entropy model algorithms are 0.3574, 0.0289, 0.6794, 0.3826 respectively. The GMM model provided the best average f-score and it is significantly better than that of the remaining three models (p<0.0001), (α =0.05), (power=80%). Conclusion: The study concluded that the GMM algorithm performed better than the other three algorithms in these four video sequences.

Article Details