The purpose of this Research work is to Detect objects in Adverse weather conditions using Dual Subnet Network and comparing it with CNN network. Object detection algorithms based on convolutional neural networks have been intensively explored and successfully implemented in numerous computer vision applications during the last half-decade. However, due to poor visibility, recognising things in rainy weather remains a considerable challenge. In this study, we introduce an unique dual-subnet network (DSNet) that can be trained end-to-end and jointly perform three tasks: visibility improvement, object categorization, and object localisation, to handle the object identification problem in the presence of fog. By incorporating two subnetworks: detection and restoration, DSNet achieves complete performance enhancement. RetinaNet is used as a backbone network (also known as a detection subnet) for learning to categorise and find objects. The restoration subnet shares feature extraction layers with the detection subnet and uses a feature recovery (FR) module to improve visibility. Our DSNet outperformed many state-of-the-art object detectors and combination models between dehazing and detection methods while maintaining a high speed, obtaining 50.84 percent mean average precision (mAP) on a synthetic foggy dataset that we composed and 41.91 percent mAP on a public natural foggy dataset (Foggy Driving dataset).