Research of improved object detection with image dehazing

Abstract

Background/Objectives: In this paper, we propose an improved object recognition method that can recover visibility by detecting degraded image due to factors such as weather, and to detect objects in the reconstructed image.Methods/Statistical analysis: In this paper, we propose improved object detection and recognition method by visibility restoration by improving degraded image. First, fog removal based on a single image is performed and restored using a median filter instead of using the minimum value in the window. We also used adaptive feature point extraction method to extract feature points by removing unnecessary information such as noise from the improved image.Findings: The performance of the proposed algorithm was evaluated by measuring the number of feature points before and after the proposed algorithm. Experimental results show that more feature points are extracted and matched than the case of recognition of the region of interest after reconstructing the image by removing the fog using the intermediate value filter in the image of degraded image by the proposed method.Improvements/Applications: Applying the proposed method to a variety of vision - based intelligent systems such as smart cars will eliminate the bad conditions such as fog, which will give better performance.