Experimental Studies of the Image Segmentation Method Quality From Unmanned Aerial Vehicles Based on the Ant Colony Optimization Algorithm Under the Influence of Additive Gaussian Noise
Year: 2025
Type: article
Abstract: The subject matter of the article is experimental studies of the image segmentation method quality from UAVs based on the Ant Colony Optimization algorithm under the influence of additive Gaussian noise. The goal is to reduce the probability of first and second type errors in image segmentation by applying a segmentation method based on the Ant Colony Optimization algorithm under the influence of additive Gaussian noise. The tasks of the study are to evaluate the robustness and accuracy of the image segmentation method based on the Ant Colony Optimization algorithm under varying levels of additive Gaussian noise, and to compare its performance with the classical Sobel filter–based segmentation approach. The methods used are digital image processing techniques, statistical analysis of segmentation quality, implementation of the Ant Colony Optimization algorithm for image segmentation, modeling of noise-contaminated conditions, and comparison of segmentation errors of the first and second kinds. The following results are obtained: the method based on the Ant Colony Optimization algorithm demonstrates superior noise resistance and maintains higher accuracy than the Sobel filter approach. Specifically, it reduces first-kind segmentation errors by 14–23% and second-kind errors by 9–17%, depending on the level of noise. Visual and quantitative analysis confirms the effectiveness of the proposed method in processing UAV-acquired imagery affected by additive Gaussian noise. Conclusions. The experimental findings confirm that the method based on the Ant Colony Optimization algorithm outperforms conventional edge detection techniques, particularly under noisy conditions, providing improved accuracy and robustness across a range of noise intensities.
Source: Advanced Information Systems
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4412839477