Speed and resolution of the most popular balance d

Balance the speed and resolution of digital camera

digital camera has two key performance parameters: image resolution in pixels and image acquisition speed or frame rate. However, the two are interrelated - the more pixels in the image, the longer the time to obtain the image, and the speed will decrease. For a specific digital camera, in order to make a proper compromise between resolution and speed, it is necessary to have a comprehensive understanding of the performance of digital image sensors and their limitations

image sensor

the core component of digital camera is image sensor, namely charge coupled device (CCD). Interlaced transmission CCD is essentially an array of pixel (photosensitive pixel) units. They convert photons into electrons and accumulate them, and then transfer them to adjacent units (vertical registers) by reading and storing the same amount of electricity as traditional batteries. The reading operation is similar to pouring all the water in a bucket into the adjacent bucket. The CCD acts as a logical shift register for accumulating charge

Figure 1 The basic CCD image sensor uses an array of shielded units (gold, forming a vertical register) to receive image data from photosensitive units or pixels (white). Then each pixel data is transported to the horizontal shift register and the output port of the camera, which is processed by the amplifier and a/d converter

the working units in CCD are connected in a chain (Fig. 1). To read a frame of image, the accumulated charge must first be transferred from a row of photosensitive units to adjacent shielding units, which form a series of vertical shift registers. The vertical register is connected to the horizontal shift register, which transports the charge to the output port of the sensor, and each clock pulse processes the charge accumulated in a unit. Several stage amplifiers and analog-to-digital (a/d) converters at the output port convert these accumulated charges into binary numbers

the correct order to read the clock signal is to first load a row of pixel data into the horizontal register, and then shift this row of data to the output port. In this way, the sensor transmits image data in row column format, one pixel at a time. At the same time, the photosensitive unit has collected the photoelectrons that form the next image

the smaller the size is, the faster the speed is.

the main factor limiting the pixel clock rate is the speed of charge transfer, which is determined by the resistance and capacitance in the circuit and varies according to the processing technology and pixel size. The general rule is that for a certain processing process, the smaller the pixel, the faster the clock speed. For existing CCD sensors, typical pixel clock rates range from 40 to 60 MHz

on the other hand, the pixel clock rate is one of the factors that determine the frame rate that the sensor can achieve. Another determinant is the resolution of the sensor, or the number of pixels. The more sensor pixels, the longer it takes to transmit the whole picture. At present, the typical sensor resolution of interlaced CCD ranges from 300000 (300K) to 16million (16m). Breaking force

it is not as simple as it seems to make a compromise between image resolution and frame rate. Other factors need to be carefully considered, including sensor structure, dynamic range and the cost of the whole camera. These factors are interrelated, which makes the balance analysis more complex

for example, the pixel size does not only affect the speed of the sensor. The smaller the pixel, the faster the charge transfer speed, but the smaller the capacity to generate and store photoelectrons. This limiting factor reduces the dynamic range of the sensor, while reducing the signal-to-noise ratio and optical sensitivity

pixel size and cost of optical elements

the smaller the pixel, the greater the influence of aberration caused by the camera lens system. As a result, small pixels have higher requirements for lens quality, thereby increasing the cost of the camera. Large pixels can often be protected from the influence of small aberration, but it makes the final size of the sensor larger. Large sensors require a larger lens to ensure imaging on the sensor, which also increases the cost. Thus, the influence of pixel size on cost forms a "bathtub" curve, which indicates that there is an optimal pixel size to minimize the cost of optical elements (Fig. 2)

Figure 2 Restricted by the performance of camera optical elements, the functional relationship between the cost of camera system and the pixel size of sensor shows an upward trend at both ends of the pixel size range

in addition, the image resolution will also affect the cost of the camera. As with large pixels, high resolution also makes the size of the sensor larger and requires a larger lens. Like other large semiconductor devices, larger sensors are more likely to have manufacturing defects. These defects reduce the output of each wafer, thereby increasing the cost of the sensor

although there is no upper limit to the number of pixels that can be accommodated by a sensor, the charge transfer efficiency of CCD units - that is, the percentage of charge transferred to adjacent units - has become a limiting factor for large sensors. If the transfer efficiency deviates from 100%, some charges will be lost in each stage of the shift register, and these residual charges will cause ghosts to appear on the subsequent picture

the transfer efficiency of CCD sensor and its efficiency of converting photons into electrons are determined by the manufacturing process of CCD. Each sensor manufacturer has its own "secret recipe" for manufacturing sensors to control the cost, size, sensitivity and speed of sensors. Therefore, the trade-offs that need to be made when selecting camera sensors vary from manufacturer to manufacturer, and are constantly changing with the progress of technology

structure improves frame rate

however, it is not only the physical characteristics of the sensor that need to be considered. Although higher resolution usually reduces the frame rate of the sensor, the reduction of the frame rate can also be compensated by the sensor structure. For example, a sensor can use multiple horizontal shift registers to divide the image into small blocks that can output data at the same time (Fig. 3). Each such block increases the frame rate, but increases the additional system cost. Each block requires an independent amplifier and a/d converter to read pixel data. In addition, these pixel reading components will be more expensive because they need to be carefully matched to maintain the uniformity of the whole picture

Figure 3 For a given pixel resolution, dividing a digital image sensor into several small blocks with independent shift registers can improve the frame rate

camera systems aimed at high-speed applications such as high-resolution photography and rocket jet dynamics research are equipped with this multi block segmentation sensor. For example, a camera manufacturer can provide an 8 megapixel camera with 16 segmented sensors and a rate of 60 frames per second (FPS). A one megapixel military camera is equipped with a sensor divided into 64 blocks, with a speed of 1000fps

appropriate tradeoffs between pixel size, resolution, frame rate and system cost are closely related to specific applications (Figure 4). The automatic video inspection system for detecting liquid crystal display (LCD) glass panels requires high resolution to detect surface defects. However, because the glass panel is difficult to move quickly, the camera does not need to be fast. On the other hand, detecting the intervals in the glass panel of plasma display has higher requirements for speed, while the requirements for resolution are less important, because there are thousands of intervals in each display. In typical machine vision applications, the speed of the camera determines the throughput of the detection station, which will affect the total manufacturing cost. For example, when the manufacturing process reaches a high tonnage, high resolution and high speed are required to quickly detect the circuit board containing small parts

Figure 4 Moving objects can be photographed at different frame rates (fps= frames per second) to improve image analysis

camera with flexible configuration

sometimes you can flexibly change the camera configuration to obtain an appropriate compromise. Handle multiple products 2 The elongation production line may need a high-resolution camera in one operation and a high-speed camera in another operation. Fortunately, flexible cameras can be achieved. This camera allows users to set it to different resolutions and speeds. For example, the camera model ipx-4m15l of imperx company can operate at a frame rate of 15fps at a resolution of 4million pixels, 30fps at a resolution of 2million pixels, or 60fps at a resolution of 1million pixels. In the dynamic production environment that needs to change the camera configuration, such a camera allows users to adjust the detection system to meet the current needs without changing the camera. (end)