Comparative Analysis of Gradient Operators in Canny Edge Detection for Bacillus sp. Microscopic Imaging

Abstract

This research aims to assess the application of microscopic imaging technology in analyzing Bacillus sp. bacteria, focusing particularly on the effectiveness of gradient operators in the Canny Edge Detection Algorithm. The study encompasses an in-depth evaluation of four principal gradient operators: Sobel, Prewitt, Roberts, and Scharr, to enhance edge detection accuracy in microscopic images. Analysis revealed that Sobel and Scharr excel in precision, with Sobel standing out in creating texture homogeneity and Scharr demonstrating superior inter-pixel correlation, both vital for ensuring visual accuracy of the images. Additionally, these operators show remarkable performance in Precision and Recall, effectively identifying relevant edges with minimal errors. Conversely, the Roberts operator, with its higher F-measure, offers an ideal equilibrium between precision and recall, making it a suitable choice for broader applications. Edge Co-Occurrence Matrix (ECM) analysis indicated that Sobel and Scharr possess higher contrast values, thereby emphasizing the sharpness of edge delineation. Conclusively, the study identifies the Scharr operator as most fitting for the analysis of microscopic bacterial imagery, owing to its capability in maintaining inter-pixel correlation and enhanced classification performance. This positions the Scharr operator as a highly applicable tool for microscopic bacterial studies, crucial in the accurate and consistent recognition of bacterial patterns. These findings significantly advance our understanding of Bacillus sp., directly impacting disease diagnostics and biotechnology. The research underscores the critical importance of selecting appropriate gradient operators in microscopic analysis and highlights the need for ongoing innovation and exploration in microscopic imaging technology.