Embedded Files
AI Full-area Patrol Gimbal
AI Full-area Fire and Disaster Prevention Patrol Gimbal
Series Solution
AI Full-area Fire and Disaster Prevention Patrol Gimbal
Series Solution
AVS uses the efficient and compact AI860 chip to construct a fire and disaster prevention smart gimbal capable of 360-degree autonomous patrol. This gimbal can be equipped with the AVS Generative GPT Edge Interactive Smart Box, providing high-level audiovisual interactive services.
The gimbal performs year-round efficient AI tracking of humanoids, vehicles, and faces, automatically monitoring restricted areas for perimeter protection. It also enhances functionality with passenger flow statistics and achieves precise identification with a facial/vehicle recognition rate of 99.9%.
The gimbal incorporates advanced FLIR thermal imaging sensors, long-range TOF laser rangefinders, and a 3D gyroscope, enhancing triple-sensor fusion. It can detect temperatures and provide fire alarms up to 500 meters for factories, 200 meters for vehicles, and 70 meters for individuals.
It is a versatile device suitable for various industrial scenarios, including manufacturing, commerce, construction, security, transportation, law enforcement, border security, unmanned aerial vehicles (drones), and other wide-ranging applications. It provides effective monitoring for occupational safety and fire prevention across these diverse environments.
A. AI AHD full area automatic patrol and fire prevention small ball camera
AVS's own AI860 chip is used to build the world's only high-definition quality AHD analog AI surveillance thermal imaging 3-light fusion full-area intelligent patrol and disaster prevention small ball camera. It can perform fire prevention, disaster prevention and theft prevention in a 75-square-foot room. It is simple, easy to use, and economical.
Functions:
Year-round 360° patrol detection, multiple preset areas, individual inspections, and full statistics.
Alarm: Send an alarm when there is an abnormality or over-temperature, including screen upload, mobile phone text message, and email notification.
Statistical chart: daily, monthly and hourly change chart.
There are also human figures, vehicles, faces, detection and tracking and peripheral restricted area warning.
Various field AI application model software will be available for upgrade in the future.
B. Audio-visual recognition and edge interactive smart box
(AI860+Raspberry Pi forms natural voice dialogue for fast video and audio interaction)
The audio-visual interactive smart box developed by AVS can be connected to the "AI full-area automatic patrol and fire prevention gimbal (see C for details)", which can further improve the accuracy of people, vehicles and objects identification to 99.9%, providing fire prevention, disaster prevention or specific site Audio-visual interactive public or emergency services. It can also be connected to the support of local GPT machines and upgraded to a generative interactive high-end smart service that is useful for various industries. The operation example is as follows:
Demo Video - Audio-visual recognition and edge interactive smart box
Demo Video - Face Recognition by AVS AI camera + Raspberry Pi
C. AI full-area automatic patrol and fire prevention gimbal (patrol range is as large as a football field)
The full-area 360° round-the-clock patrol gimbal can efficiently record, provide real-time warnings, and provide complete statistics and clear charts. It is a powerful security tool for wide-area patrols in various industries.
Specifications:
4.5 inches, 400g, long and short dual-focus full-color lens, FLIR far-infrared thermal imaging temperature measurement, long-range laser correction, 3D gyroscope stabilization calculation.
Functions:
1. Dual motors, 360° non-stop operation, multi-light day and night, full color, high definition and wide dynamic range.
2. Backlight compensation, edge enhancement, ultra-clear 2D/3D noise reduction, strong light suppression, BSI2 enhanced night vision, frame accumulation low-light brightening, red and white dual light at night.
3. Thermal image fusion 3 light zoom, coaxial 2 hundred times magnification.
4. Target tracking, full-color control, OSD annotation, video superposition, picture-in-picture multi-mode and multi-segment fusion, photoelectric calculation, high-magnification zoom and stability.
5. Multi-AI, multi-mode, and multi-model tracking and identification.
6.H.265 network day and night protection and early warning, fire prevention, disaster prevention and anti-theft automatic patrol notification.
Application fields: wafer factories, steel mills, chemical plants, electronics factories
Advanced features:
Can be combined with item B AVS audio and video recognition & interactive smart box to become a cost-effective edge smart audio and video interactive micro system.
Can add multiple objects, humanoid skeleton, high-precision license plate & face recognition comparison.
Demo Video - AI full-area automatic patrol and fire prevention gimbal
D. Thermal imaging-color vision 3-light fusion AI gimbal for drones (civilian, commercial)
Specifications and features:
1. 2 inches, 250g to 3 inches, 350g.
2. Interchangeable long and short focus full-color dual lenses.
3. 320H~640H high-end infrared thermal imaging.
4. 3D gyroscope stabilization calculation.
5. 360° dual servo motors.
6. Full color HD day and night.
7. Wide dynamic 3D noise reduction.
8. BSI2 plus frame accumulated night vision shimmer brightening.
9. Thermal imaging fusion color vision 200 times coaxial intelligent zoom.
10. AI neural network multi-processor & multi-model collaboration/humanoid/face/vehicle/object/thermal image patrol tracking real-time upload.
AVS Interactive Audio and Video Smart Box
The Ai-Pi edge AIOT system can also be used with GPT generative large models!
AI is sweeping the globe! Over the past decade, wave after wave of technological breakthroughs and applications have made AI increasingly prominent. These technologies not only gather the resources and energy required since the dawn of the electrical age but have now formed a massive AI pyramid. Particularly in the frenzied race of generative AI models like GPT, these resources are continuously concentrated. So, do AIOT (AI + IoT) systems still have a chance? Can Ai-Pi composed of AVS mini-AI and Raspberry Pi follow GPT's ascent? Today, most countries and companies can no longer afford the computational costs of large generative models, yet they are periodically overshadowed by GPT, which is hundreds of times larger. This situation leads people to believe that only through substantial centralized resource allocation can they achieve more efficient and powerful results. If this trend continues, what will the outcome be? Will AI eventually become a closed system monopolized by a few large corporations? What opportunities remain for widespread, decentralized AIOT systems?
Cloud computing touted 20 years ago, the Internet of Things (IoT) 15 years ago, AIoT 6 years ago, Metaverse 4 years ago (along with many other concepts) have all garnered significant attention in their time. These concepts were once highly anticipated. Today, while the spotlight is on large-scale GPT models in the cloud, there is also a race to develop AI smartphones and AI PCs capable of running smaller-scale GPT models. Next, can lightweight, simple, cost-effective edge AIOT systems also rise alongside GPT? What unique features and opportunities does AVS's Ai+Pi development offer?
Over the past decade, the number of global IoT devices has reached 13.5 billion, generating a market value of over trillions of dollars in applications. However, AIoT devices incorporating artificial intelligence features, such as AI computing in smartphones, still account for less than 30% of the market share. This indicates that while achieving connectivity functionality is not overly difficult, implementing effective artificial intelligence features remains quite challenging. Particularly today, there are at least thousands of chips with AI edge computing capabilities available globally (with Taiwanese manufacturers alone introducing over 200 varieties), yet only a few have achieved widespread adoption. What are the reasons behind this?
In fact, the issue lies not in hardware but in software, much like how Intel and AMD haven't fallen far behind Nvidia in graphics and gaming. However, in AI applications, they clearly lag behind Nvidia. The main reason is Nvidia's extensive experience over 15 years in scientific and engineering computing through their CUDA software development and deployment. Especially for AIOT, there's a crucial need for widespread investment in support and online software development platforms, along with the establishment of robust AI software development and application systems. Despite this, due to the broad and diversified applications of AIOT, even international giants like Google, Meta, Microsoft, and AWS struggle to provide fully adequate support. If these global giants find it challenging, what can small companies like AVS rely on to achieve success? The answer lies in the convergence of the right timing, favorable conditions, and skilled personnel----using strength and ingenuity to create miracles!
The "Yi Jing" (Book of Changes) is essentially AI and also GPT. What we refer to today as AI or even GPT generative artificial intelligence is just a small part of how the ancient "Yi Jing" manifested. It is not merely a tool for divination and fortune-telling, but requires complex and rigorous analysis and statistical consideration of many factors such as situations, environments, details of human affairs, and resources. It involves iterative reasoning over multiple parameters to deduce the most advantageous or appropriate choices. However, today's GPT still has significant room for improvement in terms of accuracy in handling multiple objectives, variables, and comprehensive behavioral assessments.
Over the past 25 years, AVS has continuously evolved its audiovisual technology and focused for the last 15 years on AI algorithms, software simulation, IC design, development tools, training systems, and extensive application development, persistently striving for excellence. Our product sales experience ranges from tens of millions to hundreds of millions of units of chips, validating our capabilities through practical applications once again. Now, we are fully leveraging our most precise, fast, compact, and energy-efficient AI860 chip system, closely integrated with the user-friendly, open, and widely adopted system of Raspberry Pi single-board computer, which has reached an installed base of 50 million units, forming the "AI Interactive Audiovisual Front-end". This system manages audiovisual interactions, including the integration of front-end vehicle, parking, and license plate recognition, and converts natural visitor speech into text for subsequent processing, interpretation, and text content adjustment.
The following content interpretation, analysis, selection, adjustment, and response are integrated through our company's local "GPT small model station platform", which consists of Nvidia GPU or AI PC. It runs the LLAMA GPT light model internally and selects the content to be returned (also filtering or adjusting content not to be displayed at the platform). When necessary, it can also connect to cloud-based backups like OpenAI or Google. Such a platform can effectively manage and serve hundreds of AVS interactive smart front-ends, forming the most cost-effective and high-quality intelligent service system for entire communities or groups (including enterprises, governments, schools, and various industries).
Application examples:
Parking lot:
"I'm not familiar with this place... Where can I park the car?"
"How much is the parking fee? Can I eat or use the restroom nearby?"
"Is there any supermarket where I can buy things?"
"what time is it now?"
"What's the weather like?"
"Help, someone is fighting!"
"Someone is setting the car on fire...Help!"
Collaborative voice chess: let your voice help you play chess
"Please help my red commander go from E0 to E1, and my black cannon needs to fly from F8 to F4."
Page updated
Report abuse