Anomaly detection, also referred to as unusual event detection, plays a critical role in video analytics by identifying abnormal occurrences within a video. However, detecting anomalies in videos presents several challenges, including difficulties in human or object identification, occlusion, clutter, illumination changes, and shape variations. To tackle these obstacles, researchers have developed machine learning and deep learning techniques, with deep learning showing particular promise. With the aid of deep learning, significant advancements have been made in anomaly detection, especially in identifying frames where unusual or abnormal events take place. The main goal of anomaly detection is to pinpoint specific moments in a video where unusual activities occur. This capability proves highly valuable in various settings, including monitoring and detecting mischievous or suspicious activities in places like shopping centers, malls, homes, banks, airports, hospitals, and academic institutions. This survey seeks to comprehensively cover the research areas related to abnormal event detection, shedding light on the progress and potential of this important field.The objective is to review and summarize the existing literature on anomaly detection in video analytics, with a focus on the latest machine learning and deep learning techniques, to identify the key challenges faced in video anomaly detection, including human and object identification, occlusion, clutter, illumination changes, and shape variations, to investigate the advancements made in deep learning-based approaches and their effectiveness in addressing the identified challenges,to explore the applications of video anomaly detection in real-world scenarios, highlighting the significance of these techniques in surveillance and security settings.