AI View Camera
AI View VA
  • The configuration needed to run the demo AI features on a server is quite large; while using a low configuration laptop you can still view the camera whilst executing other features.
  • S200: 58W (48W IR LED, IR Range 300m)
  • S201-0803F: ~14W (IR LED:3.4W, camera ~10W)
  • S500: ~40W (With high power IR LED, PT Controller).
  • S200: 33x, 300m
  • S201-0803F: 4x, 50m
  • S201-0802N: 3x, 50m
  • S500: 33x, 300m.
  • In practice, the camera can withstand wind speeds up to level 12.
  • The following PoE standards are supported:
  • PoE IEEE 802.3af
  • PoE +, IEEE 802.3at
  • PoE ++ IEEE 802.3bt.
  • S200
  • S201-0803F, S201-0802N
  • S500
  • S200: Observing highways, roundabouts, and overpasses; observing in a straight line, especially suitable for night-time surveillance at a considerable distance. Suitable for surveillance in places where maintenance is not easy. Placement is on the 2nd or 3rd floor.
  • S201-0803F: Observing highways, roundabouts, and intersections; capturing license plates to allow for the punishment of traffic violations. Placement is on the 2nd or 3rd floor.
  • S500: Observing roundabouts and; overpasses; monitoring traffic and critical locations. Placement is on the 2nd or 3rd floor.
  • S200: 350m
  • S201-0803F, S201-0802N: 50m
  • S500: 350m
  • -40 °C ~ 60 °C.
  • Yes, both features can run concurrently. Our camera has a chipset that powers AI algorithms, therefore running multiple AI features at once is perfectly fine. We have tested the simultaneous running of people counting and face recognition on the camera and experienced no performance issues.
  • Yes. Our cameras have two-layer protection for administrators, access control systems, a storage encryption system, a firewall, a VPN, and video streaming and data encryption when transmitting to the server.
  • The current processing speed is 6-7 fps. Because our current model is being continuously developed, improved, and optimized, real-time processing is completely possible.
  • We have tested our cameras in multiple environments at different temperatures, even harsh ones. However, snowy environments have not been included in trials. We’re confident the camera would function in snowy conditions.
  • Yes, it’s completely possible. Our cameras support popular protocols such as http, https, FTP, and MQTT so if your server complies with these protocols, it can be connected to the cameras. You can configure them through WebUI.
  • Bkav uses Sony and OmniVision sensors, the world eaders in image sensing solutions. Our S and P series models use Sony sensors, featuring STARVIS technology that allows viewing in low-light environments. The sensor supports a maximum speed of 60 fps, 4K resolution with IMX334, and FHD with IMX327.
  • It is possible to replace lenses on C series cameras, but you cannot replace lenses on S and P series cameras. Such changes are not encouraged, since the installation requires strict compliance with standards on electronics and mechanics. The lens assemblies used for S and P series cameras are large, with complex internal motor control mechanisms to serve capabilities like zoom, focus, etc.
  • Yes. The cameras use an anti- banding algorithm which facilitates automatic changes to the sensor's exposure time to suit the environment. Furthermore, you can manually configure the shutter speed via WebUI.
  • The auto exposure control (AEC) algorithm: exposure time is only a part of AEC algorithm. The camera can automatically change its shutter speed, digital gain, and balance between bright area gain and dark area gain to produce images with the best brightness.
  • In terms of recognition of objects and people face on; the accuracy is up to 99%.
  • Yes. Our cameras have two-way audio capability.
  • The infrared light does not affect human vision.
  • As soon as the camera detects an event, it sends information to our server. This will help minimize the amount of data on your server and significantly reduces the traffic required for that server.
  • Our cameras offer: 360-degree endless pan and; 90-degree tilt.
  • We are currently developing this capability. We will soon release and update for our cameras.
  • We have already designed PoE for this model, however the current hHousing design does not offer enough space for it. We have planned improvements for future versions.
  • Our C series cameras currently support a WiFi connection.
  • We have tools to support scanning IP address of the cameras and performing batch configuration on a single computer.
  • No.
  • Yes.
  • There’s a tool that facilitates differentiating between devices manufactured by Bkav and by other manufacturers by searching for information like the model, mask, subnet mask, and device names for devices on the same network.
  • After being installed for a long time, the cameras might get a covering of dust which might consequently cause blurring or loss of details in the images captured. The function named “Water Tank” and the wiper will help clean the glass cover of the cameras.
  • If the cameras get affected by moisture or steam, the function of the built-in heater will heat the glass cover to evaporate the moisture or steam.
  • Yes, that is possible and it is a feature on the commercial version. We are currently developing it further.
  • No, because the camera needs power to operate. If you want it to keep working, you will need an uninterruptable power supply.
  • The camera uses an adapter to convert 220V AC power to 12VDC or 24 VA.
  • Allows for computer-based observation
  • Images and audio are directly stored on athe computer
  • It is cheaper than a conventional camera system
  • It is possible to deploy multiple AI features.
  • h264 and h265.
  • We will update the features via OTA updates.
  • No.
AI feature – People count
  • ~ 10m.
  • 99% accurate - 1% miscounted, ~ 0.1% missed.
  • Yes, our cameras can count people and vehicles as well.
  • It is because some people in the frames have their human characteristics like a leg, arm, face, etc. hidden. We tried letting the camera count a person upon recognizing just one out of these characteristics, but there’s a risk that it could be taken advantage of. We are currently continuing our research and upgrades to release a complete version with an even higher accuracy in the near future.
  • RTSP.
AI feature – Face Recognition
  • ~10m.
  • 99% accurate - 1% misrecognized, ~ 10% missed.
  • ~ 4 faces.
  • No.
  • We are currently completing internal testing. We can confirm that it currently works well with 15 people in the database.
  • It uses five images from five different angles as a basis for recognizing a person within the camera database.
  • We have studied many different models and chosen the most accurate and commonly used one in the industry. The model will identify a person in front of the camera based on that person's facial features.
  • This is the first iteration of a newly developed identity recognition feature for cameras, it is not easy to convert the programming libraries from a Linux-based computer to an Android camera. Therefore, we decided to temporarily push the features with simple models and libraries to help ease the integration into a camera system with a complete AI face recognition feature. We are continuing to develop it to create a new system that requires the least input samples - one sample image – for successful recognition while reducing misrecognized cases to 0.5% like the feature that has been developed for Linux servers.
  • It can produce the most accurate recognition results when people move at a moderate speed and face towards the camera.
  • It can be used for tracking attendance in a classroom, face ID on the phone, etc.
AI feature – Vehicle count
  • In the ideal lighting and environmental conditions, our camera can work well within a distance of 300m.
  • Possible with 99% accuracy.
  • Possible with 99% accuracy.