Otimização do processamento de imagens de ovos de galinha com interface gráfica interativa e paralelismo em Python

Abstract

This work proposes the restructuring of a digital image processing system aimed at the morphological analysis of chicken eggs, with a focus on computational performance and usability. The proposal includes replacing the sequential model with parallel execu tion using the concurrent.futures library, specifically adopting the ProcessPoolExecutor class to distribute the processing of subimages across multiple CPU cores. The system also underwent a complete redesign of its graphical interface, using the Tkinter library to make user interaction more intuitive through visual buttons and organized menus. The project was structured with a modular architecture, separating the graphical interface lo gic from the processing logic, which facilitated code maintenance and enabled the efficient application of parallelism to CPU-bound tasks. The adopted methodology involved assis ted segmentation of images containing 30 eggs, as well as the automated extraction of morphological measurements, with the generation of reports and annotated images orga nized into standardized directories. Comparative tests conducted on two processors with different architectures (Intel Core i5-6200U and i3-N305) showed significant reductions in execution time, validating the effectiveness of the implemented optimizations. The re sults demonstrate that the proposed solution delivers significant gains in performance and usability, making it suitable for scenarios that require batch image analysis with greater computational efficiency