For effective Data Extraction (DE) and Data Analysis (DA), the constant flow of visual information offers unique techniques in the Video Surveillance (VS) domain. This VS application demands the significance of advanced Object Detection (OD) for obtaining Video summarization in this study. Then, the accurate detection and the location of objects in the video frames are known as OD, as it is crucial for DE in the VS. To improve OD, the research offers advanced techniques like Faster R- (Convolutional Neural Networks) CNN or FRCNN using Inception ResNet V2 (IR-V2) via the application of CNN, Region Proposal Networks (RPN) and Deep Learning (DL). The empirical outcomes indicate that this suggested framework delivers improved OD accuracy of 93.5% compared to other techniques. By overcoming Big Data (BD) in the modern VS, the combination of sophisticated Computer Vision (CV) techniques with inception modules and residual connections.