The Dragonfly AI Edge Computing Server is a self-developed, lightweight, intelligent, and edge-deployable video detection and recognition intelligent terminal. This system is designed for the video and surveillance field, with powerful AI algorithm engines pre-installed. It is a software and hardware integrated product aimed at general security applications. The synergy of AI chips, AI frameworks, and AI algorithms provides a strong and flexible AI computing support. It supports intelligent analysis of high-volume videos and enables rapid implementation of AI applications such as facial recognition, license plate recognition, people and vehicle control, boundary monitoring, and firework detection (implemented through cloud deployment). This product offers cost-effective and intelligent solutions, widely applied in various security applications such as industrial parks, campuses, construction sites, and communities.

1. Standard 1U rack server, easy to install and implement;

2. Each server supports 32 channels of video, capable of parallel recognition of 32 AI algorithms;

3. Supports the RTSP protocol, both H264/H265 are supported;

4. Can be connected to detection via NVR or camera's local area network, reducing server load and traffic costs;

5. Supports lightweight detection and alarm messages, real-time push to users;

6. Detection devices are centrally managed through cloud management, monitoring status and assigning detection tasks;

7. AI models are centrally managed in the cloud, easy to flexibly deploy and make reasonable use of resources;

8. Supports direct viewing of monitored videos, realizing risk monitoring and video surveillance interaction;

9. Built-in AI algorithm engine, combined with cloud deployment to achieve unrestricted and flexible deployment of AI algorithms for future business needs;

10. Lightweight AI model with fast computing speed;

11. Unique "edge-to-cloud" instant communication design architecture, achieving resource sharing between edge and cloud, breaking through the limitations of edge resources, and realizing full-system resource sharing;

12. Provides a rich RESTful API protocol interface, making it easy for third-party developers to integrate quickly;

13. The system supports both public cloud deployment and private cloud deployment, following business development needs.

The Dragonfly AI Edge Computing Server directly connects to network video cameras, pulls video streams within the local area network, and performs intelligent analysis and processing at the terminal. The detection results are uploaded to the cloud, reducing users' network bandwidth costs. It is flexible, convenient, and easy to deploy. The business process is as follows:

1. The detection server and the surveillance camera's NVR are on the same local area network, and video streams are pulled directly from the NVR;

2. The detection server automatically connects to the task management server, accepting management and task assignments;

3. The detection server continuously sends detection events to the message queue server for real-time monitoring by the user information platform;

4. The detection server can generate result videos and push them to the streaming media server for viewing by the user information platform;

5. The task management server provides a web platform for user operations, as well as interfaces for querying message records and retrieving violation screenshots for the user information platform;

6. The streaming media server provides distribution of detection result video streams;

7. The message queue server provides real-time distribution of detection events.

1. Smart Park: Based on the standardized access data of intelligent devices/systems in the park, combined with technologies such as artificial intelligence, big data, and computing, an integrated operation management platform for park personnel, vehicles, energy, security, assets, and environment is realized. It helps improve operational efficiency, enhance the quality of park services, elevate the level of security management in the park, and maximize the value of various intelligent systems in the park.

2. Smart Construction Site: By installing various monitoring devices at construction sites, an AI recognition and analysis system is created to build an intelligent monitoring and prevention system. This system enables comprehensive real-time monitoring of people, machinery, materials, laws, and the environment. It effectively addresses the deficiencies of traditional methods and technologies in supervision, transforming passive "supervision" into proactive "monitoring." It achieves early warning, ongoing detection, and standardized management of construction site safety, making safety production information-based.

3. Smart Community: By applying software and hardware to smart community scenarios, requirements such as facial recognition for access control, identification of strangers, detection of loitering individuals, and entrance/exit management are met. This satisfies the regulatory requirements of public security and community streets, improves community personnel management efficiency, and enhances homeowner satisfaction. It also focuses on monitoring and managing key individuals such as wanted criminals and important petitioners.

4. Safe Campus: AI-based safe campuses utilize a cloud platform, business systems, and AI face recognition intelligent hardware terminals to form a full-stack smart service. It covers applications such as big data decision-making, safety risk prevention and control, campus gate control, dormitory management systems, cafeteria consumption, cloud displays, and smart office solutions. It provides comprehensive, one-stop smart solutions for educational regulatory authorities, schools, teachers, students, and parents for creating a safe and intelligent campus.

5. Other scenarios that require the use of AI recognition and analysis.