Enhancing the low light images is a crucial task for both human vision and robotic vision. This paper presents a lightweight, hybrid approach that combine the lowlight pipeline with a Yolov8-Seg-based object detection. The system uses a global-local enhancement strategy, which adaptively enhances dark regions without disturbing the bright regions and its fine details. The post processing stage further refines contrast and colour balance which includes adaptive CLAHE (Contrast Limited Adaptive Histogram Equalization) and Gamma correction. This method is evaluated on benchmarks such as DarkFace, HRSL-Detect outperforms traditional and deep LLIE models in PSNR, LPIPS, entropy, and detection accuracy, while achieving fast runtime ( \(\widetilde{0}.99\) s) and minimal computational cost (1.215 G FLOPs). This makes HRSL-Detect suitable for deployment in edge and mobile devices.