| Title |
Particle swarm optimization-based local entropy weighted histogram equalization for infrared image enhancement |
| ID_Doc |
42032 |
| Authors |
Wan, MJ; Gu, GH; Qian, WX; Ren, K; Chen, Q; Maldague, X |
| Title |
Particle swarm optimization-based local entropy weighted histogram equalization for infrared image enhancement |
| Year |
2018 |
| Published |
|
| Abstract |
Infrared image enhancement plays a significant role in intelligent urban surveillance systems for smart city applications. Unlike existing methods only exaggerating the global contrast, we propose a particle swam optimization-based local entropy weighted histogram equalization which involves the enhancement of both local details and fore-and background contrast. First of all, a novel local entropy weighted histogram depicting the distribution of detail information is calculated based on a modified hyperbolic tangent function. Then, the histogram is divided into two parts via a threshold maximizing the mterclass variance in order to improve the contrasts of foreground and background, respectively. To avoid over-enhancement and noise amplification, double plateau thresholds of the presented histogram are formulated by means of particle swarm optimization algorithm. Lastly, each sub-image is equalized independently according to the constrained sub-local entropy weighted histogram. Comparative experiments implemented on real infrared images prove that our algorithm outperforms other stateof-the-art methods in terms of both visual and quantized evaluations. (C) 2018 Elsevier B.V. All rights reserved. |
| PDF |
|
