Using such CMOS sensors to realize digital cameras leads to the following properties: The typical sensor of modern cameras is based on CCD technology. Fig. 1 depicts a simplified description. A current of photons generates a charge which is accumulated during a certain period of time. This exposure or integration time is determined by an electronic shutter. In contrast, CMOS sensors are based on photosensitive diodes which are connected to resistors in series (Fig. 1). As a result, the photo current is continuously converted in to an output voltage. |
![]() |
| Features Using such CMOS sensors to realize digital cameras leads to the following properties: Random access: Since the conversion process is continuous it is possible to realize cameras whose pixels are randomly accessible in any sequence and at any moment (see fig 2). Thus we can think of such a camera as a kind of RAM. Once an address (in the form of a row and column index) is fed into the camera, the corresponding data (graylevel) is immediately outputted. Thus the connection between the application software and the sensor is very close. Common CCD technology, on the other hand, is based on a video standard (e.g. CCIR or EIA) which forces us to handle a stream of 50 (EIA 60) fields per second in the form of an analog signal. Thus, the connection between the application software and the sensor is very loose. |
![]() |
| High frame rate: Random access of pixels means that the frame rate is determined by the frequency of the pixel clock and the number of pixels which define the frame (which is freely definable due to random access). With a pixel clock of 16 MHz and a frame of 512 by 256 pixel we obtain 120 full frames per second. When lacolizing the frame to 128 by 128 pixel we obtain 1000 images per second.
Very high dynamic: Due to special features of the CMOS structure the conversion process is non-linear. The CMOS sensors used in LOGLUX cameras show a dynamic of much more than 6 decades (The sensor are calibrated at 9 decades.). Low power consumption: The power consumption of CMOS cameras is about 1 decade lower than that of CCD cameras. Noise: The straightforward realization of a CMOS sensor (as illustrated in fig 1(b)) would result in an excellent noise source whose raw image would be almost useless because the achilles heel of CMOS sensors is the problem of matching the multiple different amplifiers within each sensor. Fortunately, this problem has been overcome, reducing the residual level of fixed-pattern noise to insignificant proportions. Spectral Sensitivity: From the esthetical point of view, images acquired by logarithmically responding sensors may look strange, but they are ideally suited to metrology and surveillance applications. Furthermore, CMOS sensors do in general show another behaviour as CCD sensors regarding their spectral sensitivity. The following diagram shows the spectral sensitivity of the HDRC4 sensor. |
|
![]() |
|
| If you require more detailed information please have a look at our download pages. | |