Image Enhancement for DOM data production under ENVI

Multi-spectral image color synthesis methods can be divided into natural true color synthesis and non-natural pseudo color synthesis. Natural true color synthesis means that the color of the synthesized color image is similar to or consistent with that of the actual image, the general method is the synthesis of the red, green, and blue of the multi-spectral image corresponding to R/G/B; the non-natural false color is the opposite.

Natural true-color synthesis of remote sensing images can be divided into the following methods: 1. Synthesis of multi-spectral images using red, green, and blue channels is generally used for high-resolution images; 2. Weighted processing using other bands, generate red, green, and blue CIDR blocks again, which are generally used to enhance the color layers of certain features, such as vegetation and water bodies. 3. Use other band information to regenerate a certain band, it is generally used for sensors without band, such as spot and Aster.

1)
Band weighting (including R, G, B, and NIR band data)

The most common method is to enhance vegetation information. Green and near-infrared band weighting can be used. The formula is as follows:

Bandnew = A * bgree + (1-A) * bnir

A is the weight value, 0 ~ 1.

Bandmath and layerstacking tools under ENVI are used below to synthesize real-color images using TM images.

(1) Enable TM images including R, G, B, and NIR bands.

(2) Select basic tools-> bandmath, input expression: byte (B2*0.8 + B4 * 0.2), and select the corresponding band for calculation.

(3) In the band list, select R-red band, G-synthetic band, and B-blue band.

(4)
Basictools-> layerstacking, which combines three bands into a three-band image file.

In addition, in order to make the color of the land outside the vegetation more authentic, only the vegetation can be enhanced. In this case, the difference between vegetation and vegetation can be distinguished by the difference between vegetation and vegetation. First, calculate the difference value first, and use the following band expression to enhance the computation:

(B3 GT 0.2) * (B2*0.8 + B4 * 0.2) + (B3 le 0.2) * B2

B3: Beijing.

2) band generation

The common method is to generate the blue band of the SPOT image. The following is a common method of the SPOT Image:

(1) The original green wave segment (0.50-0.59 μm) is treated as a blue wave segment (which is close to the spectral range of the blue wave segment), and the red wave segment (0.61-0.68 μm) the original band is still used, and the green wave segment is replaced by the arithmetic average values of the green wave segment, red wave segment, and near-infrared band.

R: xs2

G :( xs1 + xs2 + xs3)/3, bandmath expression: byte (fix (B1) + B2 + B3)/3)

B: xs1

The effect is as follows:

 

(2) regard the original green wave segment (0.50-0.59 μm) as the blue wave segment, and the red wave segment (0.61-0.68 μm) still uses the original band, green Wave segments are replaced by green wave segments and near-infrared bands by a weighted average value.

R: xs2

G :( xs1 × 3 + xs3)/4, bandmath expression :( B1 * 3 + B3)/4

B: xs1

The effect is as follows:

 

The above two algorithms are similar, so the color effects of generating images are not much different.

The above method is not limited to SPOT images, but can also be used for ASTER images, and even for other data including green, red, and near-infrared bands.