EmbedTek Invention Improves Video Surveillance Capabilities
EmbedTek filed a provisional patent that fulfills a need, in particular, for video surveillance applications, where in many cases, a relatively small portion of the image is considered an area of interest.
The provisional patent is for a non-linear, progressive lens constructed to optically capture an area of interest positioned within the 3D field of view for the camera. A camera system with a progressive lens may be used to monitor objects (e.g., vehicles, people, people counting, safety, security, etc.) that are in the desired zone of the field of view for many different applications (e.g., monitoring a roadway at a tollbooth, license plate recognition, monitoring a storage facility, etc.).
Existing camera lenses have a uniform distribution of pixels. They are unable to provide the desired resolution of objects that are both close to the camera and those farther away within a defined area of interest. At the same time, many pixels that are processed are not used because they identify objects outside the area of interest. The EmbedTek progressive lens solves these issues by redistributing pixel density to form a unified pixel density within a desired zone 3D space. Desired resolution can be achieved and wasted pixels are avoided.
The invention consists of:
- A camera
- A non-linear, progressive lens constructed to have a unique profile (e.g., contour, shape, and/or size) designed for each specific 3D application
- One or more camera sensors with an aspect ratio defined by the total number of pixels in the X-Y plane
- A processor that records and analyzes a final image
The desired zone is predetermined by the application to which the system will be applied. As a result, the construction of the progressive lens may vary based on the placement of the camera or the desired zone of the image the camera is monitoring. The desired zone may include disjointed or unconnected portions of the area to be monitored.
The following example shows how the system can improve video surveillance in a tollbooth/roadway application. The field of view in this example includes a continuous section of roadway taken by a tollbooth camera as well as areas around the roadway.
Existing Cameras: Camera lenses with a uniform distribution of pixels
Figure 1 illustrates an image taken by a typical camera with objects that are in close proximity to and far away from the camera. The image captured by the camera has a field of view that includes vehicles that are fairly clearly visible in the shallow area, and vehicles that are blurred and obscured from clear view.
Existing cameras cannot separate out or only focus on the desired area or frame, which in this case is the roadway. They cannot eliminate or disregard the undesired areas such as the tree line on each side of the roadway, as well as the skyline. Also, the uniform pixel distribution of the camera lens causes the objects that are positioned relatively close to the camera, inside and outside the frame, to appear with a higher resolution than is necessary to view objects in those areas. At the same time, objects relatively far from the camera appear with a lower resolution than is necessary to adequately capture and process relevant objects.
This image, which was taken by positioning the existing camera above the roadway to monitor vehicles on the roadway and identify specific vehicles via their respective license plates, does not adequately provide the ability to readily identify vehicle A and vehicle B in the same image. This is shown in the schematic illustration on the right, where the license plate of a closer vehicle is easily readable on the image, but the license plate of a vehicle located farther away cannot be read or discerned. The vertical scale of the image is not linear, so the second vehicle and the second license plate appear much smaller than the first vehicle and the first license plate.
The Progressive Lens Difference: Pixels captured within a desired 3D geometric zone
Figure 2 illustrates how in the same video surveillance scenario above, a progressive lens camera system would only focus on the roadway. The progressive lens is constructed to optically capture and monitor the roadway without capturing the undesired zones of the tree line and sky. The desired zone is bounded by a frame that includes a continuous section of the roadway with the first vehicle having the first license plate and the second vehicle having the second license plate.
The progressive lens camera system is installed and positioned relative to the roadway. It redirects pixels outside the area of interest both vertically and horizontally and, consequently, the pixels that are positioned within the frame, to define the final image that fits or fills the aspect ratio of the camera sensor.
In this example, the progressive lens expands an end of the frame horizontally and vertically, and sides horizontally, to form the adjusted frame.
Expanding the frame increases the size of the second vehicle and, in turn, the second license plate so that the license plate is the same size or approximately the same size as the first license plate. As a result, the first license plate and the second license plate, which are in different areas of the zone, are both visible and readable for monitoring of the vehicles.
EmbedTek is currently applying its progressive lens technology as part of its design and development of complex vision systems, including video surveillance. The engineering team continues to develop new system designs and non-linear lenses for a wide range of applications to remap each sensor pixel to be best utilized within a targeted area. This allows the use of less expensive sensors to capture better resolution within targeted areas.