Because the particular color of the image is reconstructed from the combination of the values received from the three different colors reflected, variations can not only cause errors, but can cause amplified errors if there are variations in different directions between a green and blue LED, for instance.
解答例
Another problem with devices of this type is that the photodetectors themselves have different responsivities depending upon the wavelength of light received.
Thus, for the same radiant flux of light reflected for two different colors, the photosensor would provide an output which indicates a difference in value and thus induce an error in the color of the image detected.
FIG. 4 illustrates this effect for three different LED types.
FIG. 4 is a graph of wavelength along the horizontal axis and relative spectral response along the vertical axis.
A solid line 28 indicates the sensor responsivity of the photosensor.
As can be seen, this varies depending upon the wavelength, with lower responsivity for lower wavelengths.
Superimposed is an illustration of the wavelength distribution for LEDs of three particular color designations.
Any individual LED will emit many different wavelengths of light whose radiant flux is roughly in accordance with the distribution indicated.
As can be calculated using the photopic response function 26, a first LED type 20 (blue) has a relative spectral power distribution where, for the same integrated luminous flux as provided in the relative spectral power distribution of a second LED type 22 (green), the photosensor would provide an output for the first type that is roughly three times that of the second type.
Similarly, a third LED type 24 (red) has a relative spectral power distribution where, for the same integrated luminous flux as provided in the relative spectral power distribution of a second LED type 22 (green), the photosensor would provide an output for the third type that is roughly 12 times that of the second type.
Another effect is illustrated in FIG. 5 showing the luminous flux output as a function of electrical power for three different LED types.
A first LED type 1-1, labelled 20 (blue) has a much lower output luminous flux that the other two LED types.
LED type 1-2, labelled 22 (green) has nearly the same luminous output flux as LED type 1-3, labelled 24 (red/orange).
Accordingly, in addition to the photosensor having different spectral responses, the output of the LEDs themselves will vary for the same amount of power applied to the LEDs.
There have been basically two approaches to compensating for the above identified effects.
In a first method, such as set forth in Sharp U.S. Pat. No. 4,731,661, compensation is achieved by varying the amount of time the different light emitters are on.
In the Sharp patent, three different fluorescent tubes are used for three different colors.
The amount of time each is on is varied to compensate for varying spectral response of the photosensors.
The photosensors typically used respond to sensed light with an output voltage which is an integration function of the received light.
Thus, the longer the light is on, the higher the photosensor output.
Thus, having a longer on-time for a wavelength to which the photosensor is less sensitive will compensate for the lesser sensitivity.
Another example of how compensation is done is set forth in U.S. Pat. No. 5,268,752.
In this patent, LEDs of different colors are used.
For each LED, the output of the detector is compensated for with a compensator circuit.
U.S. Pat. No. 4,833,533 sets forth in more detail how a compensation of the photosensor output can be done.
In the '533 patent, a sample strip is read with a sensor, and a value is stored for each pixel for each color.
In other words, the compensation is applied on a pixel by pixel basis and on a color by color basis for each pixel.
This compensation value is stored in the memory, and then used to calculate a compensating ratio which is applied when the scanner is used to read an image.
The present invention provides a combination of first providing a coarse adjustment by varying both the power provided to the emitters and the duration of the emitter/detector exposure time, and then providing a fine adjustment by storing a correction value for each photodetector element.
