Electronics-How CMOS image sensors work

Both charge- coupled- device (CCD) and complimentary- metal- oxide- semiconductor (CMOS) image sensors work on the same principle. They use photodiodes to convert light into electric charges, which are then stored and processed digitally. Reconstruction from the digital data yields the photograph. The sensors contain millions of photodiodes forming an array. To record color images, each imaging pixel contains four photodiodes, one for red color, one for blue color and two for green colors. This is achieved by placing a color filter on top of the respective photodiode blocking all other colors, since the photodiodes themselves are color insensitive.

The difference between CCD and CMOS sensors is that each pixel in a CMOS sensor has its own addressing and amplifying electronic circuits, while all pixels in a CCD sensor share the same amplifying circuit. While CCD sensors have dominated the market historically, CMOS sensors start to take over recently, especially in high-end products.

CCD Image Sensor. A typical CCD image sensor is shown in Figure 1. The main body is the dense array of photodiodes with color filters on top. A single pixel contains four photodiodes dedicated to red (1), green (1) and blue (2) as shown in the magnified view. Beneath each color filter is a photodiode that converts light into electric charges. Electrons generated are stored in a photodiode potential well and subsequently transferred across the chip through registers and output to an amplifier. The amount of charges generated is proportional to the light intensity. The schematic diagram illustrated in Figure 1 shows various components that comprise a typical CCD image sensor using full-frame architecture.


Figure 1, Schematic of a CCD image sensor using full-frame architecture. The array of imaging pixels with color filters ontop form the main part of the sensor. Alongside the chip are electronics needed to read the stored charges of each pixel.

The image sensor is placed at the focal point of the lens in a camera or camcorder. All of the photodiodes in the array collectively act as the image plane and are available for detecting photons (light). Upon exposure, each pixel converts the light (Red, Green, Blue) at the point into electric charges. Lights of different intensity generate different amount of charges. These charges are then shifted downward in a parallel fashion, one row at a time, to the serial CCD shift register (illustrated as a series of gray-scale elements at the bottom of the pixel array), starting from the row immediately next to it. This process is electronically controlled by the parallel shift register clock control unit. After that, they are then shifted leftwards one at a time to the output amplifier (controlled by the serial shift register clock control unit), followed by an analog-digital-converter (ADC) to save the information digitally. All charges must be clocked out of the serial register before the next row of image data can be transferred to the shift register. This process is repeated until all rows of image data have gone through the output amplifier and ADC and stored in a memory. Reconstruction from the stored digital data yields the photograph. The conveyer belt model shown in Figure 2 is wonderful illustration of this process.

About jlw

Junling got his PhD in Materials Engineering from University of Maryland, College Park. He is currently a Professor at Nanyang Technological University, Singapore, teaching undergraduate courses related to Quantum Mechanics.