Soil Classification using Image Processing

Published on 03 Jul 2021. Written by Palvi Soni

Classifying soil is in great demand as it helps to investigate the site and provides relevant knowledge about the materials that support construction there. Soils are classified based on different properties, it can be on the basis of location and on the basis of the size of particles in it. We can classify soil here on the basis of location and texture. Due to the rise in inflation hiring skilled labours and workers incur high cost. Conventional method of soil classification like pressure meter test, vane shear test are somewhat time- consuming and tricky as well. There is a dire need of automation in this field as well so we in this project illustrate about the usage of image processing techniques for classifying soil.

Read more..

SLNOTE


Technologies used for this project:

MATLAB 2019 a: MATLAB 2019 a has some new features that better supports in classifying various kinds of the images, as it has default image processing box inside it.

Proposed Methodology for Classifying soil

1. Gathering Soil Images
2. Storing images in the database
3. Preprocessing of soil image
4. Perform feature extraction and selection
5. Apply the segmentation technique
6. Apply classification technique

SLLATEST

• Gathering Soil Images: For performing this task we need images of soil, we can either capture these from various regions and stored into a database or we can get these from online sources. Criteria are images used should be colored.
• Storing images in the database: After gathering images we need to store it on the database.
• Preprocessing of soil image: To remove the noise or distortions from the images we need to apply some image enhancement methods. It is mainly done to achieve a good quality of the images. Color maps are used here for providing better quality to edges of the image.
• Perform feature extraction and selection: Feature extraction better known as variable extraction basically done to extract important features of the images such as texture, color, quantity etc. We extract vital variables on the basis of which further process will be taken forward. After extracting feature suitable feature for classifying soil is selected. Soil can be classified on the basis of color as well as the texture of the soil
• Apply segmentation technique: We can use various segmentation techniques to perform this task but here we make use of K-means clustering technique to reach a step near towards the solution.
• Apply classification technique: Classification is the process of categorization, we apply some classification algorithm to classify different types of soil.SVM can be used to solve this problem, but KNN approach can give more accurate results. KNN is k-nearest neighbour, it predicts values on the basis of its k-points. This algorithm not only helps us to improve accuracy but it generates results in less time.
• Detection of soil type: After applying KNN classification, we can easily detect the type of soil based on its texture and location.

Advantages of this project:

1. It increases efficiency
2. It saves human effort
3. It saves time
4. Overhead cost is diminished.
5. It generates more accurate results

SLDYK
Kit required to develop Soil Classification using Image Processing:

• No kit required

Technologies you will learn by working on Soil Classification using Image Processing:

• MATLAB
• Image Processing

Any Questions?