Testing Oncam Grandeye's Panoramic Cameras (IPC)Author: John Honovich, Published on Dec 20, 2009
For almost a decade, panoramic surveillance cameras have been promised as a 'next big thing' but has yet to meet its potential.
In that last few years, however, the situation is changing as megapixel and IP video systems become increasingly main stream - enabling easier, lower cost development and deployment of panoramic cameras.
Grandeye was one of the most well known panoramic camera manufacturers but suffered a major company restructuring one year ago raising questions on its future. In early 2009, OnCam Global announced that it acquired the "worldwide rights to Grandeye technology and products for security applications."
Adding to the hype for panoramic cameras is Mobotix's intense marketing campaign for their own panoramic cameras.
A few key findings from the test include:
- A clear tradeoff exists between panoramic megapixel and 'traditional' tele-photo megapixel cameras: Much greater FoV for significantly less details
- Fairly large areas can be monitored for activity with a single camera in the center of the space
- Integration with Milestone worked well with a few notable issues/limitations
- Low light and WDR conditions were problematic
Key Recommendations and Findings
Since using panoramic cameras are a new and significantly different experience than other cameras, including 'traditional' megapixel, we emphasize the following findings:
- The Halocam is a megapixel camera but unlike 'traditional' megapixel cameras, the Halocam provides broad situational awareness but not high detail forensics - this is a fundamental tradeoff between concentrating pixels in a relatively small area (like a megapixel camera with a telephoto lens) vs. a 360 degree area
- The camera's visual quality looks like CIF video with the major benefit that it is CIF video in all directions; it feels rough equivalent to having 6 - 8 cameras streaming at CIF next to each
- In small areas (say under 400 sqft or 40 sqmeters), a 'traditional' megapixel camera with a wide angle lens placed in a corner will likely provide equally good coverage at a lower cost than the Oncam
- In mid-size areas (approximately 2000 sqft or 200 sqmeters), placing an Oncam in the center of that area is likely better than using 'traditional' MP or SD cameras - simply because multiple megapixel cameras would be needed to cover an area of this size
- The camera will not capture minute details (like facial features or license plates) farther than 5 or 10 feet from the camera.
- Warning on High Ceiling Camera Mounting: Our tests in the garage where done using a 9 feet ceiling - a fairly common height in deployments. However, use with higher ceiling heights (15 and 20 feet heights often exist in production) will reduce the effective image quality/details captured of people at floor level. This is a simple result of the Halocam using a fixed length focal lens. By contrast, traditional cameras with varifocal lenses can simply have their focal length adjusted to optically 'zoom' in closer to people. Specifiers should carefully consider this and evaluate placing Halocam cameras on extension mounts for high ceilings.
- At more than 10 feet from the camera, clear facial details are not captured; in constant daytime lighting, people can be visually identified at 100 feet (and perhaps farther); however, at lower light conditions (even 10 lux), visibility of 50 feet can be difficult
- Wide variances in lighting can be especially problematic as the camera covers such a wide field of view (in different directions). Since it's a single imager and with a single exposure, it's not possible to maximize exposure when one side is extremely bright and the other much darker. This is more challenging than using 2 separate physical cameras where each camera's exposure can be optimized for its own narrower Field of View.
- Low lighting video quality tends to be dark and moderately noisy but consistent with 'traditional' surveillance cameras that do not have a mechanical cut filter.
- Very low light conditions (under 1 lux) could be quite problematic as the image becomes both dark and noisy. The performance is consistent with 'traditional' cameras lacking a mechanical cut filter.
- On two slides of the imager, a semi-circle lip obscures objects in the upper half of the FoV. The camera can be physically rotated and installed to place those sides viewing less important areas. However, this should be kept in mind during installation.
- Frame rate for the 4 MP full fisheye stream was routinely in the 3-6 fps range as reported by Milestone and confirmed by our visual verification. While frame rate can depend on a wide variety of factors, we felt the visible frame rate was sufficient for general monitoring.
- While the camera supports multiple modes (ceiling, wall and table top mounted), the most commonly used and likely most effective use case is ceiling mounted. We wall mounted the camera and it worked fine but a lot of the FoV was wasted looking at the ceiling and the floor.
- We did light testing of the motion detection/tracking. While it seemed to work reliably in well, evenly lite scenarios, continuous false alarms were triggered in low or wide dynamic range lighting conditions.
While our test is of the ceiling mounted version (the IPC), all cameras in the lineup have a 5MP sensor, a 360 degree fisheye lens, a maximum 4MP stream and the same fundamental software functionalities.
Key software functionalities include:
- Offers 5 simultaneous streams: the full 4MP fisheye and up to 4 virtual cameras (VCAMS) that provide up user definable subsections up to 1024 x 768 resolution.
- Motion detection and tracking that allows an object to be identified and the object to be tracked across the entire fisheye FoV using the virtual cameras
- Dewarping software that allows users to pan, tilt and zoom 'immersively' throughout the entire Field of View
The MSRP of the IPC is $1995 USD. The lens is built in. Licenses for 3rd party VMS software is sold separately.
Use and Configuration Review
While configuration was fairly simple, using panoramic cameras feels different. Keep in mind the following:
- You can use the full fisheye view alone, the VCAMs alone (1 or up to 4) or a combination of both. We focused the test on using the fisheye view as it presents different benefits and challenges than using VCAMS (which are more like traditional cameras).
- When integrating with Milestone, we had 1 configuration problem with setting up the PTZ controls. In the setup, under fisheye settings, the default is unchecked for ceiling mode. Since ceiling mode is the default and likely best use for this camera, you have to check it (This is shown in the screencast below).
- The camera does not have 'traditional' exposure optimization settings (like adjusting the maximum exposure time to specific values). Instead, it has two options (flicker and motion blur modes) that indirectly impact exposure. We explain this in the screencast.
- The user's manual is short on details (at only 28 pages) and can leave the integrator with important unanswered questions.
The screencast provides an overview of the physical size and form factor of the IPC tested. There was not significant issues or surprising elements. Note that the unit does not have an external network activity LED which can make troubleshooting difficult as it can be tricky to determine if the camera is powered on or not.
Image Quality Analysis and Samples
In the screencast below, we examine the video quality with emphasize on:
- Image details: because the Field of View is so wide, the image is not very detailed (compared to other megapixel cameras). However, by contrast, it provides modest details across a far wider area.
- Wide Dynamic Range conditions: In our test of a parking garage that is partially outdoors, we demonstrate tradeoffs in configuring and using the camera for wide dynamic situations.
- Low Light scenarios: In low light, the camera becomes darker and noisier, consistent with color cameras without a mechanical cut filter.
Download of Video Samples
The 7 video sample of the Oncam IPC may be downloaded (NOTE: 220 MB zip file).
Competitive Comparison and Recommendations
The Halocam competes against 3 categories of products: (1) traditional megapixel, (2) non-immersive panoramic cameras and other immersive panoramic cameras. The following is summary recommendations on each category:
- Against traditional Megapixel: If you want a small area of very high details (like a doorway or or an entrance), then a traditional megapixel camera is a better choice than the Halocam. However, if you want to cover a moderately large area with a moderate amount of details, then the Halocam can provide greater value. We believe it is better to consider the Halocam as the replacement for multiple standard definition cameras in a confined area than as a direct competitor to megapixel cameras.
- Against Non-Immersive Panoramic Cameras: ArecontVision offers a 360 degree camera that combines (4) 2MP imagers into a single housing. The cameras are displayed as (4) separate video feeds placed side by side. It is not immersive and has a blind spot directly below the mounting spot of the camera. However, on the positive side, it does not require dewarping software (because it consists of 4 'traditional' megapixel feeds) and it has a significantly higher rated pixel count. It already enjoys fairly wide 3rd party VMS support (see support document) and has an online price of about $1600, roughly similar or less than Oncam's $1995 MSRP. In November 2009, Avigilon announced plans for a similar non-immersive 8MP 360 degree camera.
- Against Other Immersive Panoramic Cameras: Two other notable suppliers of immersive panoramic cameras include Mobotix and Sentry 360. While the technology is fundamentally similar (immersive video through dewarping of video from a fisheye lens), Mobotix market approach is different - much lower cost (about $1000 for the panoramic camera and free VMS software from Mobotix) that is more targeted at the home/SMB market. Sentry 360 is a more direct competitor to Oncam as both seek to integrate their panoramic cameras with 3rd party VMS systems. (See our Sentry 360 overview). Beyond this, until we test Mobotix and Sentry 360, we cannot judge relative quality.
Most Recent Industry Reports
The world's leading video surveillance information source, IPVM provides the best reporting, testing and training for 10,000+ members globally. Dedicated to independent and objective information, we uniquely refuse any and all advertisements, sponsorship and consulting from manufacturers.