Night sight object identifier using HSV color space for lighting invariant object detection

Object detection in night-time environments remains challenging due to low illumination, noise, and loss of visual details. Most deep learning–based approaches require large datasets and high computational resources, limiting their suitability for real-time and low-cost systems. This paper proposes an adaptive HSV-based night vision object detection framework that enhances low-light images using Contrast Limited Adaptive Histogram Equalization (CLAHE) and performs object detection through adaptive thresholding and morphological analysis. The proposed method operates without deep learning models, enabling fast and interpretable detection in real-time. Experimental results using webcam-based input demonstrate improved visibility, effective foreground segmentation, distance of the object from a camera and reliable object localization under varying night-time lighting conditions. The system is suitable for driver assistance, surveillance, and embedded vision applications.