Crowd analysis is gaining a lot of attention in both the technological and social aspects as the population expands. As certain crowdactivity cannot be assessed easily, These techniques differ dependingon the crowd. Strikes, protests, and other crowded events are commonin overpopulated cities. Parades and other public gatherings providea variety of security challenges. Security forces are frequently enlistedto keep an eye on these issues and ensuring the safety of those whoattend meetings. In these circumstances, ensuring the safety of thosewho attend meetings are challenging . Cutting-edge crowd control methods are necessary for an intelligent surveillance system. The goal of theproposed method is to find crowd analysis subfields, such as anomalydetection, that have not yet been well explored or that seem to beaddressed infrequently. Public data sets for indoor and outdoor useare used to evaluate the proposed system’s bespoke architecture. The Experimental findings demonstrate that the custom architecture performs better than the exixting Architectures.Our aim is to providerecent state-of-the-art algorithms to provide a novel strategy that canhelp with improved crowd surveillance and crowd activity analysis.