Інтелектуальна система виявлення близькорозташованих малорозмірних об’єктів: архітектура та реалізація

Abstract

The article analyzes the development of an image formation simulation model for optoelectronic systems and a com-bined method for detecting small, low-contrast objects in digital images. The proposed model enables the generation of synthetic datasets that account for physical and technical factors affecting image quality, providing a valuable tool for testing image processing algorithms. Additionally, the detection method, based on an adapted MobileNet V1 architec-ture, demonstrates high accuracy and robustness to noise under limited computational resources, making it suitable for embedded applications.