Dragonfly AI Camera is the industry's first low-power solar AI camera. It supports 4G communication and the GB28181 protocol. It is a self-developed lightweight, intelligent, and flexibly deployable AI camera for video streaming, detection, and recognition. This system is aimed at the video and surveillance fields, is preconfigured with a powerful AI algorithm engine, flexibly deploys various recognition algorithms from the cloud, and is positioned as a software-hardware integrated product for broad security applications. AI chip + AI framework + AI algorithm work together to provide strong and flexible AI computing support. It supports solar power supply and 4G network deployment, quickly enabling AI applications such as face recognition, license plate recognition, people and vehicle control, boundary warning, and fire and smoke alarms (implemented through cloud deployment), providing users with flexible intelligent solutions. It is widely used in various broad security application scenarios such as parks, campuses, construction sites, and communities.

1. It can use solar power and 4G network deployment, and can also use mains power and broadband deployment, allowing flexible installation and implementation according to application scenarios;

2. Supports H264/H265 streaming;

3. Supports lightweight detection alarm messages pushed to users in real time;

4. Cameras are uniformly connected to cloud management for status viewing and detection task assignment;

5. AI models are uniformly managed in the cloud for convenient and flexible allocation and rational use of resources;

6. Supports direct viewing of surveillance video, enabling interaction between risk monitoring and video surveillance;

7. Supports viewing playback of recent surveillance recordings;

8. Built-in Dragonfly AI algorithm engine, working with the cloud to realize unlimited and flexible deployment of AI algorithms for any future business requirements;

9. Lightweight AI models with fast computing speed;

10. Unique “edge-cloud” instant interconnection design architecture, realizing resource sharing between the edge and the cloud, breaking through edge-side resource limitations, and achieving system-wide resource sharing;

11. Provides rich RESTful API protocol interfaces for third-party integration development and rapid business integration;

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

Dragonfly AI Camera features AI functional head technology. It combines an AI head and a machine, adopts machine capabilities, and through automated technology, reasonable analysis of scenes, machine recognition, image recognition, and intelligent visual learning, completes models.

Main Hardware Functions

1. Edge Computing Capability

Programmed AI chip with computing capability, completing processing capability at the intelligent end.

2. Low-Power Design

This system design improves solar power and facilitates the key power supply underneath.

3. Image Options

Adopts a video sensor standard, reaching 1080; main stream video encoding dynamic night scene does not conform to H.264; main stream 1080P, sub-stream 720P, sub-stream 540P; does not conform to 1920*1080.

4. Front-End Hardware Storage

The device video storage capacity is ≥6 hours, supporting no data loss.

5. Data Transmission

Can transmit video images in real time and supports telecom operators' 4G networks.

6. Ambient Temperature

Normal operating temperature: -20℃～+60℃.

7. Night Vision Function

Starlight-level night vision, supports uploading clear images.

8. Faults

Has excellent equipment and functions.

9. Stability

Under normal operating conditions, it can work continuously for 30 days or more without electrical, mechanical, or software failures.

10. External Power Supply

Has an external power interface, with an optional 220V external power supply.

11. Photovoltaics and Battery

Power supply lasts 3 days in rainy weather; separate structural design from the camera; the batter

y lithium battery adopts lightweight installation and continuous installation, reducing and maintenance costs.

12. AI Recognition Capability

(1) Has direct AI recognition function: the front end can perform recognition (including basic and data front-end recognition) (including basic and data front end), without transmitting back for recognition;

(2) Intelligent recognition: recognition can be performed without uploading images;

(3) Recognition: identifies the minimum requirement for 0 radiation degree of the environment.

(4) Can recognize targets with risks of people, vehicles, fire, and smoke, with a recognition rate ≥90%; missed detection rate ≤0.3%. Currently target support is required; correctness can be based on different recognition methods according to different needs and degrees;

(5) When network conditions are poor or the network is disconnected, the recognition results will not be affected by network interference.

Dragonfly AI Camera is widely used for personnel intrusion prevention warnings in key locations such as villa courtyards, factory warehouses, warehouse material custody, military border warning area intrusion, unattended power stations, high-speed rail platform passenger line crossing, highway entrances and exits, pedestrian zebra crossings, railway crossings, perimeter fences along railway lines, prisons and detention centers, airport perimeters, oil pipelines, military restricted zones, industrial and mining enterprises, residential communities, homes, campus walls, agricultural anti-theft, building property management, government premises, oil depots, subways, railways, tunnels, scenic areas, ports and wharves, railway stations, bank intrusion warnings, and museum perimeters.

Personnel climbing over and scaling walls recognition warnings; detection and warning of personnel illegally entering restricted areas; perimeter intrusion prevention detection and recognition warnings; recognition alarms for personnel entering gates; recognition alarms for personnel intrusion in factory production areas; detection and recognition alarms for personnel crossing boundaries, etc., all fall within the prevention scope of Dragonfly AI Camera.

Application Scenario 1: Smart Construction Site

Ensures personnel safety in engineering work areas. Engineering safety control capabilities support dynamic detection such as face recognition, safety helmet detection, reflective vest detection, gathering/fighting, fire and smoke detection, fire and smoke recognition, personnel intrusion, and phone use.

Application Scenario 2: Smart Park

Through Dragonfly AI Camera, for industrial parks and other locations, it monitors personnel climbing behavior; monitors key areas based on electronic fences to prevent personnel intrusion; performs fire and smoke detection in key areas, improving park safety and reducing manual patrol costs.

Application Scenario 3: Smart Chemical Industrial Park

Meets the integrated control needs of chemical enterprises across safety, environmental protection, emergency response, and production, such as off-post/leave-post detection, smoking detection, PPE wearing detection, open flame detection, leakage detection, personnel intrusion, and personnel falls.

Application Scenario 4: Smart Power

An edge-side product based on intelligent recognition services for power plant security and intelligent management scenarios. It can analyze and share video content in real time, automatically discover potential risks, and proactively issue risk warnings.

Application Scenario 5: Smart Factory

Smart factories further standardize manufacturing processes and workflows, improve product quality, increase production efficiency, effectively reduce violations of safety protection requirements, and greatly reduce safety accidents and personnel injuries.

Application Scenario 6: Smart City Management

Through Dragonfly AI Camera, for streets and main roads, it performs visual recognition of excessive gatherings and fighting; monitors illegal and random parking of motor vehicles and non-motor vehicles; and monitors street sanitation.

Application Scenario 7: Smart Environmental Protection

With Dragonfly AI Camera, monitoring is carried out on key streets, areas, and other regions to integrate human society and environmental operations and build a better city.

Application Scenario 8: Line Protection

Through Dragonfly AI Camera, it collects and compares the environment around utility pole lines to check whether the surroundings and lines are normal;

For example: 1. abnormal attachments and damage by animals and plants; 2. unauthorized connection and illegal construction of lines; 3. unexpected line problems such as line detachment; 4. real-time alarm prediction for fire, smoke, etc.; 5. utility pole tilting; 6. gathered crowds, possibly indicating human construction; 7. overtime and overfrequency vibration, possibly indicating mechanical construction or natural disasters.

Through computing by Dragonfly AI Camera, enterprises can promptly identify the source of problems, reduce losses and the number and severity of damage, and trace the causes of accidents, thereby establishing sound response measures and ensuring the physical safety of communication facilities.

Application Scenario 9: Worker Safety in Production

Through Dragonfly AI Camera, it performs personnel identification for workers, detects whether workers correctly wear safety helmets and whether they cross lines (electronic fences) or climb, and performs fire and smoke detection in key areas, helping enterprises fully and promptly understand production conditions and strengthen safety production management and supervision.

Application Scenario 10: Behavior Analysis

Through Dragonfly AI Camera, it analyzes and processes abnormal personnel behavior. It can be applied to machine recognition of behaviors such as key area prevention, surveillance of important items, and leaving suspicious dangerous objects. It can also alarm abnormal personnel behavior, greatly improving the application efficiency of video surveillance. In addition, it can implement crowd situation analysis, such as crowd density analysis and personnel gathering analysis, analyze the situation in key areas or places with large gatherings, prevent crowd incidents, and achieve early warning and timely handling.

Application Scenario 11: More Scenarios

Various suitable scene recognition algorithms can be embedded according to project needs.

Q1: What power supply and networking methods does the Dragonfly AI Camera support?

A: It supports solar + 4G wireless deployment with no wiring, and can also connect to 220V mains power and broadband network cables. The solar power system has a split design and can maintain continuous operation for about 3 days even in rainy weather, making it suitable for field locations.

Q2: What are the characteristics of the Dragonfly Camera's edge computing AI recognition?

A: AI recognition is completed locally on the device without uploading images to the cloud. Even when the network is disconnected or poor, the functions of recognizing people, vehicles, fire, and smoke still operate normally, and recognition accuracy can reach over 90%.

Q3: Which protocols does the Dragonfly AI Camera support, and is it easy to integrate with third-party platforms?

A: It natively supports the GB28181 protocol and provides RESTful API interfaces. Both public cloud and private cloud deployment modes are supported, making it convenient to integrate with a project's existing security or business management systems.

Q4: What scenario-based AI detection alarms can the Dragonfly AI Camera implement?

A: It can perform perimeter intrusion, fire and smoke recognition, safety helmet detection, personnel boundary crossing, smoking detection, etc. Different algorithms can be flexibly delivered from the cloud, and it can be used in construction sites, chemical industrial parks, and power lines.

Q5: Can the Dragonfly AI Camera work stably in harsh environments?

A: The enclosure has an IP66 protection rating and an operating temperature of -20℃~+60℃. It features starlight-level night vision and stores recordings locally on an SD card. Even if the cloud connection is disconnected, local recording data will not be lost.