Return
Enhanced dark soil image processing: A comprehensive algorithm for quality improvement
DOI:10.1016/j.optlastec.2026.115354.png)
Abstract
En 中文
The quality of soil images is very important for automated soil analysis in farming, environmental monitoring, and land management. Nevertheless, soil images taken under low light or poor illumination conditions frequently suffer from noise, weak texture visibility, and low contrast, which adversely influence the performance of computer vision and AI-based soil classification systems. To overcome these limitations, this paper presents a hybrid framework for image enhancement named as Tri-Enhance Soil Algorithm (TESA). It integrates Anisotropic Diffusion (AD) to suppress noise while preserving essential structural details, Dynamic Stochastic Resonance (DSR) to boost weak soil texture characteristics in dark regions, and subsequently, Contrast Limited Adaptive Histogram Equalization (CLAHE) to improve contrast in specific areas. When these strategies are used one after the other, they can effectively reduce noise while also making edges clearer, textures more visible, and contrast better in dark soil pictures. Numerous experiments on various low-light and dirt pictures show that TESA consistently outperforms traditional enhancement approaches in both visual quality and quantitative evaluation measures. In addition, TESA-enhanced images improve visual clarity and structural separability of soil textures, which is beneficial for downstream soil analysis tasks such as automated classification. The suggested framework serves as a strong and useful preprocessing step for data-driven soil analysis and precision agriculture applications.
Keywords:
Anisotropic diffusion
Dynamic stochastic resonance
Contrast limited adaptive histogram equalization
Image enhancement
Journal
O
IF:
5
Papers:
2.4K
Citations:
3.5W
Organization
Cited Papers
No cited papers available

