Determining Camera Specifications and Specification Evaluation Methods
How much detail can a camera reproduce?
Monochrome Cameras
How well the subject can be recognised in the dark in night-time monitoring can be identified by looking at the minimum subject illumination specifications of the camera.
Image recognition products best known for their use in vehicle license plate reading systems have become widely used in recent years. When the video signal sent from the camera is decoded through software, a large amount of noise in the video signal may significantly hinder the software’s ability to decode the signal. Therefore, it is essential to look at the S/N ratio of the camera prior to use so that the noise levels of the output signals can be taken into consideration.
This page briefly explains approaches and methods used in determining camera performance and specifications, and related evaluation methods defined by Ikegami. The users of our cameras demand high resolution, high sensitivity, and an excellent S/N ratio. In order to continue providing products of the highest quality, Ikegami determines the specifications for these most important basic functions based on the Ikegami standards, which contains strict rules and guidelines for company manufacturing standards.
The performance of this function depends on the horizontal resolution, and each manufacturer has different specifications. So, how does Ikegami determine horizontal resolution?
Method for Determining Specifications
Horizontal resolution is determined by taking the same width as the vertical screen size but in the horizontal direction, and finding how many monochrome vertical lines can be recognised within that area.
Horizontal resolution is basically decided by the number of pixels of the image sensor in monochrome cameras. As the aspect ratio of NTSC or PAL systems is 4:3, the resolution is calculated as the number of horizontal pixels multiplied by 3/4.
Examples:
ICD-48E 752 pixels x ¾ = 564 TV lines Official value 560TV lines
ICD-48 768 pixels x ¾ = 576 TV lines Official value 570TV lines
Colour Cameras
Colour cameras are equipped with colour filters for the image sensor, and an optical low pass filter, which is to be mounted on the front of the camera. The effect of these filters results in a value that is 85-95% of a monochrome camera with the same number of pixels.
Method of Measurement
An image is actually projected onto a resolution chart, and visual measurement is carried out on the monitor screen.
Method for Determining Specifications
Minimum illumination can be specified by identifying how light conditions need to be to obtain a recognisable image, and by determining the lowest level of light required to obtain a certain image level. As this will change with the F value (*1) of the lens, the F value must also be stipulated. Ikegami uses F1.4 as a standard, but more recently has also used F1.2.
Important Image Level
The stipulation of image level becomes important at this time. If the image level is too low, it becomes impossible to recognise the subject on the screen. For this reason, Ikegami sets the image level of its colour cameras to 50% of the normal state, and monochrome cameras to 25% (or 50%) of the normal state. Furthermore, Ikegami will state the specified image levels in future catalogues.
Example: 0.5Lx/F1.4 at 50% (0.5Lx/F1.4 at 50IRE)
Method of Measurement
Measurements are taken by projecting a grey-scale chart, attaching an ND filter to the front of the lens, and lowering the light level uniformly.
(*1) F value: This is a number that represents lens brightness; the smaller number, the brighter lens. The relationship between brightness (F value), focal length (fl) and effective diameter (D) is described by the following equation: F=fl/D.
Method for Determining Specifications
The ratio between the signal level and level of noise contained in the video signals is shown. In general, a level of 48dB or more is adequate for obtaining a picture quality suitable for monitoring use.
Method of Measurement
Camera settings such as AUTO GAIN CONTROL (AGC) and detail compensation are set to "OFF" and a white chart is shot and measured. As regular CCTV cameras have a gamma property of 0.45 (*1), the signal level is adjusted to 60% so that the inclination of the gamma curve becomes 1, and the value is read using a noise meter. If an oscilloscope or wave monitor is used for measuring, it is necessary to convert the noise level into an effective value.
The Characteristics of Human Eyes
As fine random noise is difficult to capture as opposed to large frequent noise that is easy to capture, one’s perception of image quality differs depending on the frequency components or noise cycles, even if the S/N ratios are the same. Point-like high-frequency noise is relatively inconspicuous, but low-frequency horizontal noise is very conspicuous, and therefore becomes a major factor in the degradation of image quality.