版本 1.0.3 (2.7 MB) 作者: Ravi Kakaiya
We are analyzing and classifying the Land type using texture and color analysis.

64.0 次下载

更新时间 2021/11/21

来自 GitHub

在 GitHub 上查看许可证


We are analyzing and classifying the Land type using texture and color analysis.

Color analysis

From different images we can observe the color intensity mean values. and Can find the dominant color present. This will help to separate between some class like barren land vs. Forest. Here shown the Mean values of intensities for Different type of land images:

Color\Land Agricultural Barren Forest Houses Mountains
Blue 86 145 49 123 116
Green 81 158 53 126 132
Red 64 180 38 133 138

As from above table we observe that Genrally for forest land the Green part is dominant.

Texture Analysis

Properties\Land type Mountains Houses Forest Barren Agricultura
Contras 0.56 0.25 0.69 0.059 0.114
Correlation 0.92 0.96 0.79 0.92 0.904
Energy 0.05 0.08 0.08 0.44 0.32
Homogeneity 0.78 0.88 0.74 0.97 0.94

Entropy value

Land type Mountains Houses Forest Barren Agricultura
Entropy 7.91 7.77 5.71 6.17 6.41

Approach 1(Individual image):

-> Here we start with, if the value for entropy is higher than some threshold we can say there are higher probability of the land containing Mountains.
-> Than Secondly we find the highest present color in the image ,if it comes out to be Green than we can say the land have forest.
-> We calculated Gray level co occurrence matrix(GLCM) and from that we found different properties(Contras,Correlation,Energy,Homogeneity)
-> Now we see that the lowest value occurs for contrast than it is equal chances for Barren land or aggricultural land.
-> Than we see if the dominant value of color is red than very likely the land is barren land. Else it is aggricultural land.
-> The Rest land can be treated as land containing Houses.

Approach 2(Multi image):

-> Here we pass the images having different type of land view.
-> Than we Find the Gray level co occurrence matrix(GLCM) for all the images.
-> Using GLCM we Found the properties of the Texture (Contras,Correlation,Energy,Homogeneity)
-> Here as we have values for all type of land so, we can compare in between them, which can give a better result.
-> Here as for Mountains, the shadows and bright parts form a drastic change in intesity so it have highest value of contrast.
-> For the Houses we observe that correlation is very high because it show the inter relation between pixel to pixel.
-> Now For Barren land as there will be uniformity in the image so we observe high value of Homogeneity.
-> The remaining land we can classify as Aggricultiral land because it can contain mix properties of Barren land as well as Forest .


Ravi Kakaiya (2022). Land-Classification-Based-on-Color-and-Texture (, GitHub. 检索来源 .

MATLAB 版本兼容性
创建方式 R2021b
Windows macOS Linux
标签 添加标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!
要查看或报告此 GitHub 附加功能的问题,请访问其 GitHub 仓库
要查看或报告此 GitHub 附加功能的问题,请访问其 GitHub 仓库