Mainly by means of infrared nmled 940, 940 nm semiconductor laser, infrared lamp source, such as all kinds of infrared light source excitation (
, reflected light through the 940 - nm filter to eliminate or filter out the stray light, ensure that the sensor (
Received signal accurate, don't come out the phenomenon such as distortion or saturated and garbled.
Large amounts of data acquisition and transmission.
Light source characteristics of face recognition of image acquisition device, light source -
- - - - - -
Using high power infrared diode, in the majority with 850 nm and 940 nm wavelength.
In order to improve the recognition efficiency and improve the utilization rate of light from the light source selection began to consider the overall design.
Although on the market to buy the LED nominal value is 850 nm and 940 nm, but in the measurement of specific LED products center wavelength, it found that there are still many deviation.
In 850 nm LED, for example, the actual center wavelength of 835 nm, 865 nm.
By F light source used in face recognition system for high power LED array, if each LED the center wavelength of * no, after all LED spectrum superposition, narrow-band filter selection and matters needing attention narrow-band filter is mainly used to isolate the interference light, through the signal light, fully highlighted the useful information, reduce the interference information, and lay a foundation for subsequent image processing and recognition.
At present, face recognition is mainly used in various occasions, and entrance guard system.
Have a plenty of installed in the indoor light dark places, have a plenty of installed in a brighter place.
Different occasions, the interference light intensity is different, so the requirement of narrow-band filter is also different.
Narrowband filter of narrowband filter bandwidth is not as narrow as possible, also is not as wide as possible, according to the environment and decide together the light source.
850 nm infrared LED its bandwidth around 50 nm, choose narrow-band filter were should consider the utilization rate of light energy, so you can't set the narrowband filter bandwidth too narrow, the LED light source, the bandwidth of the under 15 nm is not appropriate.
On the one hand, because is too narrow bandwidth will LED a large part of the stronger signal light, the other party and because is too narrow the bandwidth of the filter is small, the effective use of Angle may cause taken to bright dark edges in the image.
Tested and found to leds luminous intensity threshold can be used in about 70%, the image of the still has quite a good contrast, as a result, the bandwidth of the narrowband filter can choose around 30 nm, require a slightly higher against interference, can choose 20 nm bandwidth.
Two kinds of main 850 nm narrowband filter, one kind is 30 nm bandwidth, the other is a 20 nm bandwidth, the people who choose more.
Narrowband filter cutoff range narrow-band filter cut-off range is mainly according to the response of the receiver itself range and receiver interference sources in the environment determined wavelength range.
Receiver CCD response range is 400 ~ 1100 nm.
In face recognition, interference sources mainly diffuse reflection or scattering sunlight and surrounding artificial light, cross the wavelength range is very wide, delay zhong from ultraviolet to near infrared.
Integrated the two reasons, face recognition using narrowband filter cutoff range can be determined for 400 ~ 1100 nm.
Narrowband filter application in face recognition face recognition technology is to the person's facial features information identification, it is a kind of biological recognition technology.
Using image sampling device contains a human face image or video stream, and according to the image automatically detecting and tracking human face, and the face feature location, extraction, by comparing identification to identify different people as to the identity of the day.
Face recognition arithmetic is very huge, and the initial image quality and algorithm of knowledge, don't have a decisive influence efficiency.
Here, we mainly for image acquisition device is used in face recognition system of narrow-band filter is analyzed, the mouth is to help users better understand the function and using method of narrowband filter, so that the correct selection of the technical index of the narrowband filter.