Dehazing is a challenging big task for the professionals who are trying to emerge in image processing. The objective of video dehazing is to process a blur, hazy video frame by frame. Hence, we make image more clearly and when compared to actual video. The objective of this paper is to get clear, dehazed video from hazy, blur videos which will be used for different important real-time applications like surveillance, crime detection, etc. The technology used for video dehazing and object identification is deep learning, a dynamic science which incorporates with machine learning. Dark Channel Prior algorithm generates a dehazed video using a far and near convolution neural network and deep learning models like YOLO (You Only Look Once), SSD (Single-Shot Detector) are widely used. A succession of experiments is further more accomplished to exhibit. The algorithm that has been suggested has the ability to generate haze-free images of exceptional quality, showcasing a plethora of well-defined details, minimal color distortion, and negligible halo artifacts. It outperforms or is comparable to four state-of-the-art haze removal algorithms. This achievement is due to advanced techniques that enhance image clarity while preserving natural colors, ensuring visually appealing and accurate results.