D-ISDET: Double Intensity of Image Shadow Detection and Elimination in Autonomous Vehicle

Abstract

We nominate the double intensity of image shadow detection and elimination method which is called D-ISDET. It offers to restore the invisible image information in the image which is covered by the shadows. The D-ISDET which supports the certainty of object detection in the autonomous vehicle is running well. Hence, the disparity of the shadowed space and the true non-shadowed space can be distinguished with using the double intensity and threshold methods. Likewise, in the big problem of imprecise detection of low-shadows and the perplexed shadows, the double intensity method is endorsed to reinforce the true shadow spaces, thereby the threshold method abolishes the shadow correctly. The experimental results of D-ISDET performance indicate the enhancement of the BER and RMSE indexes for shadow detection and elimination, respectively. D-ISDET outshines achievement between 0.24% and 1.85% of the shadow area detection and between 1.42% and 3.05% of the shadow-free area detection compared to the other methods. D-ISDET also works out between 4.11% and 16.59% of the shadow elimination and D-ISDET reaches between 0.60% and 16.57% of shadow-free elimination compared to the other methods. D-ISDET also carries out the first-rate performance compared with the other methods.