Car Lane and River Bank Detection through Ensemble Learning

In this study, we propose a novel approach to enhance edge detection for car lane and river bank detection through ensemble learning. We developed a refined version of the UNet model, named UNet refinement, and combined it with Restormer and ZoomNet models, each trained in different environments, to predict edge images. The predictions were integrated using a voting-based ensemble learning method. This study highlights the effectiveness of ensemble learning and our UNet refinement module in improving the performance of individual models for complex image processing tasks related to car lane and river bank detection.