People that are perpetually hunting for knowledge will benefit from data acquisition. The early phase in video data acquisition is splitting the video into images. Many images are tiny and don't reveal a lot about the picture's information. Scene boundary identification, or video segmentation into action sequences, enables a more complete comprehension of the image sequence by classifying images based on comparable visual content. The purpose of this article is to discuss video scene recognition, particularly video structure extraction for pattern comparison with significant properties. The article designed and developed a methodology that would include stages for image collection, detecting commonalities among frames, selecting important frames, and detecting the time at where the relevant frame is identified. The pictures are generated by Python's OpenCV and scene classification metrics are used to assess the method. As assessed by numerous parameters, the findings shows that scene identification and accuracy are considerable. Furthermore, we investigated and researched contemporary identification and assessment techniques. Moreover, we have tested extensively our research framework on a variety of publicly available event video databases, and these outperformed several futuristic techniques. The outcomes of this research can be utilized to generate real-time definitional video assessments.