Scribd
Upload
52 views1 page

Analysis of Fast DCP Algorithm: 5.1.1 Down-Sampling

The Fast DCP algorithm aims to overcome the main drawback of long computational time in the original DCP algorithm for dehazing images. It does this through three main steps: 1) down-sampling the input hazy image to reduce spatial frequency, 2) estimating the dark channel image using only single pixel patches instead of large 15x15 patches to reduce computation, and 3) estimating ambient light to further reduce computation time and improve performance over the original DCP algorithm.

Uploaded by

swethachand7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
52 views1 page

Analysis of Fast DCP Algorithm: 5.1.1 Down-Sampling

The Fast DCP algorithm aims to overcome the main drawback of long computational time in the original DCP algorithm for dehazing images. It does this through three main steps: 1) down-sampling the input hazy image to reduce spatial frequency, 2) estimating the dark channel image using only single pixel patches instead of large 15x15 patches to reduce computation, and 3) estimating ambient light to further reduce computation time and improve performance over the original DCP algorithm.

Uploaded by

swethachand7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 1

Analysis of Fast DCP Algorithm

The main drawback in DCP algorithm is it’s computational time. Practical applications
require less time to calculate the dehazed image from hazy image. In order to overcome this
drawback we use an improved version of DCP algorithm which is a Fast DCP algorithm.

In the above section 4 we mentioned complete analysis of DCP algorithm [1] in which the
' '
transmission medium is t ( x) is calculated by using local patch. So, the edge region of t ( x)
is rigid. In order to smoothen these edges we used Transmission Map Refinement, which is a
time consuming process and can’t applicable for practical applications. We reduce this effect by
using Fast DCP algorithm.

This Fast DCP algorithm mainly contains three steps as follows:

1.Down-Sampling
2.Estimation of Pixel-wise estimation Dark Channel Image
3.Estimation of Robust ambient light

5.1.1 Down-sampling

From He’s method [1] we observe that dark channel of haze image I has low spatial
frequency. So, we calculated dark channel value with a down-sampled image inspired by [5].

5.1.2 Estimation of pixel-wise Dark Channel Image

We calculated the dark channel value using one pixel (1×1 patch) instead of 15×15 patch
[1] to reduce the computation time without losing the spatial information. Hence, our proposed
method do not need soft-matting process.

5.1.3 Estimation of Robust ambient light

In order to reduce the computational time and to improve

You might also like