Section 2 concludes that a glare meter based on a digital camera is probably too expensive to develop and produce, and may not be simple in use.

Transcription

1 Possible development of a simple glare meter Kai Sørensen, 17 September 2012 Introduction, summary and conclusion Disability glare is sometimes a problem in road traffic situations such as: - at road works at night with glare caused by flashing warning lights or lighting of the work zone - in night traffic with glare caused by road lighting, luminous/illuminated advertisement signs or road signals. Therefore, the question has been raised within the NMF (Nordic Meeting For improved road equipment) if it is possible to arrive at a simple glare meter in the sense that it is relatively cheap to develop/produce and easy to use. Disability glare is generally subject to calculations in connection with design calculations of lighting installations in particular road and tunnel lighting installations. Calculations are based on a glare formula that was developed a long time ago on the basis of experiments. This formula is introduced in section 1 and it is explained that disability glare is caused by scattering in the human eye and is expressed by a veiling luminance. According to the glare formula, the contribution to the veiling luminance from a glare source varies with the angle from the line of sight to the direction towards the glare source. Therefore, a glare meter must provide the correct variation of the sensitivity with a sufficiently good accuracy. Further, a glare meter must be calibrated directly into the unit of veiling luminance. The sections 2, 3 and 4 explore the possibilities to provide the variation of the sensitivity by means of respectively a digital camera, non-sharp imaging or blurring of sharp images. Section 2 concludes that a glare meter based on a digital camera is probably too expensive to develop and produce, and may not be simple in use. The two last-mentioned approaches are based on a single detector that measures the veiling luminance in an integral manner. Such glare meters are probably simple to use. The detector needs to be a photodiode, which has a large range of measurement, a good linearity and probably allows for measurement in all relevant circumstances. The photodiode needs a reasonably good spectral correction in terms of overlaying filter glasses. Further, there has to be a suitable amplifier and a simple panel to show the results. The panel needs to have a range change in order to cover a large range of the signal. Finally, a set of cross-hairs or a sighting scope is needed for aiming the instrument. Section 3 concludes that non-sharp imaging must be based on a non-standard lens, which is designed, tested and produced for the purpose. This adds to the cost of development and production. An example of a design is given. Note: The expensive Pritchard photometer, supplied by the company Photo Research, once had a glare lens as an option. This seems not to be the case anymore.

2 Section 4 concludes, on the other hand, that blurring of sharp images can be based on standard optical components. However, the fine tuning of the sensitivity requires experiments that add to the cost of development and possibly to the cost of production. Accordingly, it is possible to design a glare meter with relatively simple optics that measures the veiling luminance in an integral manner by either of the two approaches of non-sharp imaging or blurring of sharp images. However, the optics need to be developed and tested and its is necessary to consider the detector, its spectral correction, an amplifier, a panel, means to aim the instrument and the physical construction. Two other matters that are not considered in the note, should be mentioned as well. One matter is calibration that needs to be done with a calibrated light source. It would be necessary to rely on a calibration service in a similar manner as with lux- and luminance meters - probably with repetition of the calibration each or every second year. The other matter is time varying lights, such as flashing warning lights. These introduce a need for logics that extract effective values of the veiling luminance, which poses a complication in itself. The overall conclusion is that a glare meter may be easy to use, but cannot be really cheap to develop and produce. 1. The glare formula The human eye is not quite clear, and for this reason, illumination on the eye causes an illumination by scattering on the retina like a veil of light. This is called disability glare as the veil of light overlays reduces the contrasts of objects within the scene of view. The strength of the veil is described by a veiling luminance, which in itself is a measure of disability glare. However, in applications it is necessary to set the veiling luminance in relation to the luminance level within the scene of view. As an example, the characteristic used in road lighting the threshold increment TI is to some approximation the ratio between the veiling luminance and the road surface luminance. There is a single formula for deriving the veiling luminance in calculations, which is of almost general use: Lv = K E/θ 2 where Lv is the veiling luminance caused by a single light source K is a constant E is the illuminance on the observers eyes caused by the glare source and θ is the angular separation between the observers line of sight and the glare source. The following comments apply: - Lv is measured in the unit of luminance of cd/m 2 - the value K = 10 is used for young observers; the literature provides typical increasing values with age - the illuminance on the observers eyes is measured in lux (lx) - the angle θ is measured in degrees. The formula is normally considered to be valid for a range of θ from 2 to 20, but is often used in a larger range. Some other formulae are sometimes used to provide even larger ranges. When more than one glare source contributes to glare, the total veiling luminance is obtained as the sum of the contributions from the individual glare sources. When a glare source has a large angular subtense, its contribution to the veiling luminance can be obtained by integration over the angular subtense. 2

3 Figure 1 illustrates the angular variation of the sensitivity to glare, while figure 2 shows the variation of the term 10/θ 2 of the glare formula. Figure 1: Variation of the sensitivity to glare. Veiling luminance as a function of the angle to the glare source 3,0 2,5 Figure 2: The veiling luminance as a function of the angle to the glare source (θ). Veiling luminance 2,0 1,5 1,0 0,5 This variation forms the basis for the development of a glare meter that measures the veiling 0, Angle to the glare source (degrees) 3

4 luminance in an integral manner. The glare meter must provide the correct variation of the sensitivity with a sufficiently good accuracy and must be calibrated directly into the unit of veiling luminance. The variation is obviously strong with a factor of 100 for the range of θ from 2 to 20. For this reason it is not easy to design a glare meter. 2. Use of a digital camera It is obvious to base a glare meter on a digital camera. The principle is that an image of the visual field is summed up with weights in accordance with the glare formula. The image itself may serve as documentation for the geometrical circumstances of the measurement. In practice there will be some difficulties: a. The camera will be subject to vignetting from the lens and, therefore, show a variation of the sensitivity over the image. This variation has to be measured and compensated for. b. The high luminance of glare sources in combination with a relatively poor range of sensitivity of a digital camera will introduce the need to set the aperture of the camera in individual cases in order to avoid saturation. Alternatively, neutral filters can be used. c. The rather poor pixel resolution of a digital camera may introduce the need of blurring the image so that small glare sources are made to cover a number of pixels. d. Because of b and c, an image that shows the glare sources will mostly not show the visual field to a useful degree and cannot after all serve as documentation. e. The vignetting of a camera depends on the setting of the aperture of the camera, so that different compensations are needed. f. A digital camera of a good quality is an expensive component. g. There will be a lot of programming. Because of these difficulties, a glare meter based on a digital camera may not be cheap and easy to use. 3. Non-sharp imaging It is also obvious to use optics like in a luminance meter, but out of focus to as to broaden the measured field and cause variation of the sensitivity. Figure 3 illustrates the optics of a luminance meter, where a sensitive surface of a detector is focussed sharply on a measured field by a lens. Figure 4, on the other hand, shows blurring of the measured field by defocussing. Figure 3: The optics of a luminance meter. 4

5 Figure 4: A blurred field with a defocussed luminance meter. Blurring can also be introduced by the use of a lens that is not suitable for sharp imaging all over the lens surface; for instance by the use of parabolic, elliptic or even hyperbolic lenses. However, detailed calculations reveal that the correct variation of sensitivity cannot be obtained by this kind of blurring, neither by defocussing, lens type or a mixture of the two. It is in fact necessary to design a lens for the purpose, and one such is shown in figure 5. The lens has a diameter of 70 mm, a plane lower surface and an upper surface with a curve that changes gradually from downwards at the centre to upwards at the periphery. The lens is designed for a small detector (less than 2 mm diameter), a position of the detector in the axis of the lens 60 mm below the plane lower surface of the lens and a refractive index of the medium of the lens of 1,5. The lens collects light within a cone of ± 20 with a sensitivity in accordance with the glare formula and transmits it to the detector in a cone of ± mm Figure 5: The cross-section of a lens. 5

6 Before producing lenses like that, it may be necessary to produce a test lens of a sufficient quality to verify the performance. Additionally, the lens should be designed for a larger detector in order to provide more signal. This, and the actual production of such lenses, may be expensive. 4. Blurring of sharp images The common way of shaping the luminous intensity distribution of a signal light is first to produce a quasi parallel beam by means of a parabola or a lens, and then to spread the light out from the beam with a spreading front glass designed for the purpose. It is, therefore, obvious to return to the simple optics of a luminance meter, but think of blurring by means of a spreading glass in front of the lens. Attempts to design a such a glass for the case of glare were made, but failed. It seems that internal reflections in the patterned surface of a spreading glass prevent the strong variation of sensitivity that is needed in the case of glare. Note: The design calculations were made with a ray tracing program that includes reflections, but does not account for them in other ways than the overall result. Instead, the optics illustrated in figure 6 has been designed. The optics include the following components: - a lens with a diameter of 55 mm, a plane lower surface and an upper surface with a parabolic shape providing a focal point in the axis of the lens placed 58 mm below the plane lower surface of the lens and a refractive index of the medium of the lens of 1,5 - a 50 mm disc of a 3 mm thick opal plate that transmits/reflects 60/40 % of the incident light and is placed with the centre of its upper surface in the focal point of the lens - a detector placed with the centre of its receiving surface 4 mm below focal point of the lens and a diameter of the receiving surface of 2 mm - a 50 mm disc of a mirror reflector placed with the centre of its upper surface 10 mm below the focal point of the lens - a circular 10 mm black field in the centre of the reflector. The sensitivity of the optics has been calculated at 2, 3, The match to the correct curve is good as shown in figure 7. The type of lens, the opal plate material and the reflector material are standard products. The tricks of obtaining a good match is to select the various properties of the optics, such as the type of the opal plate material, the size of the black field and the depth positions of the opal plate, the detector and the reflector. In practice it is necessary to develop a prototype, where the match is obtained by trial and error. 6

7 Figure 6: Optics using a parabolic lens, an opal plate and a reflector with a black field. Approximation to the glare formula by optics 3,0 2,5 2,0 glare formula approximation by optics Figure 7: Approximation to the glare formula by the optics. Veiling luminance 1,5 1,0 0,5 0, Angle to the glare source (degrees) 7