WO2008119405A1

WO2008119405A1 - Prevention of halos around intense light sources for optical systems, particularly for night vision systems

Info

Publication number: WO2008119405A1
Application number: PCT/EP2008/000626
Authority: WO
Inventors: Jörg Moisel
Original assignee: Daimler Ag
Priority date: 2007-03-29
Filing date: 2008-01-28
Publication date: 2008-10-09
Also published as: DE102007015599A1

Abstract

The invention relates to the problem of the formation of halos (glare) around intense light sources, particularly for night vision assistants in motor vehicles. To solve this problem, the invention proposes a special intensity reducer, e.g., a filter (9), which causes a strong drop in the point spread function of the objective.

Description

Avoiding halos around intense light sources in optical systems, especially in night vision systems

The invention relates to the problem of eliminating halos (glare phenomena) around intense light sources, such as in night vision assistants in motor vehicles (for example the "NightView" system from DaimlerChrysler)

Headlights of oncoming vehicles in the one of

Camera captured image it can occur from a camera system.

A halo is generally an unwanted effect, which is also referred to as a "halo" around very bright image objects around and can cause glare. This problem of halos generally occurs in imaging systems in which an image is taken with a camera that includes a lens and an imager.

The problem of halation occurs in particular with images taken with night-vision systems, for example with a night-vision device based on the principle of residual light amplification, with or without illumination of the recorded scene. The invention will hereinafter be considered to be the preferred one without restriction of generality Application example of a night vision system of a motor vehicle explained.

In such a night vision system, also referred to as night vision assistants, a scene is taken with a camera and the image is displayed on a monitor for the driver. As a rule, such pictures are kept in black and white. The night-vision system is more sensitive than the human eye, so the driver can use the night-vision system to view things, e.g. People, obstacles or sources of danger that he would otherwise recognize worse, not or only much later.

On the images displayed by the monitor, however, not only are the dark objects that are not visible to the eye alone displayed, but also bright objects such as intense light sources. In such intense light sources, which may in particular be headlights oncoming vehicles, circular halos can be seen around the light sources on the monitor. These halos are not merely a blemish of the system, but they seem to magnify the imaged objects, thus contributing to glare and reduction in resolution, and thus are a phenomenon limiting the safety and usefulness of the night vision system.

Although a high development effort for the cameras and in particular for the camera lenses is operated, it has not been possible in the prior art to avoid these Halos or eliminate them with little effort. It has previously been assumed that the cause of these halos is remaining in the lens manufacturing shortcomings (Eg surface roughness) is that can not be eliminated or only with disproportionate effort.

From the document DE 10 2004 028 616 A1, a camera lens with a wavelength-dependent f-number is known, wherein the wavelength-dependent F-number is set by the radial and spectral profile of the transmissivity of at least one filter in the diaphragm.

Document DE 10 2004 030 661 A1 discloses an optical low-pass filter comprising a plurality of light-conducting optical fibers.

From the document DE 699 01 677 T2 an optical anti-aliasing filter for use in optical systems is known, which realizes a low-pass filtering for spatial frequencies.

The document DE 100 17 185 A1 describes a camera with a non-linear dependence of an electrical output signal that can be generated by it on a brightness that can be received by the camera, in order to adapt the brightness dynamics and / or the contrast of the recorded image.

The invention has for its object to reduce the formation of halos, especially in night vision systems, or to avoid, with a low in practical application technical effort, without consuming or lengthy digital image processing methods are required.

The object is achieved by a camera lens with the features of claim 1. Preferred embodiments and further developments of the invention result from the dependent claims and the description below with associated drawings.

A camera lens according to the invention, in particular for a night vision system with a camera, preferably with an infrared-sensitive CMOS camera in a motor vehicle, which has a diffraction-limited point spreading function as a function of the lateral coordinate with respect to a direction perpendicular to the optical axis of the objective. Function), thus shows the peculiarity that the lens comprises an optical intensity attenuator arranged in the imaging beam path which attenuates the light intensity in the beam path of the objective in a lateral direction according to a predetermined pupil function whose Fourier transform decreases with increasing lateral coordinate (x) in that, starting at a specific value x ', it is smaller than the diffraction-limited polychromatic point spread function of the objective without the intensity attenuator. In this case, the polychromatic point spread function of the objective corresponds to the envelope of the monochrome point spread functions or corresponds to the smeared different monochrome point spread functions. An exemplary possibility is given by the fact that the Fourier transform of the predetermined pupil function decreases more rapidly with increasing lateral coordinate (x) than the diffraction-limited polychromatic point spread function of the objective without the intensity attenuator.

This makes it possible to modify the optical properties of the objective so that the influence of portions of the point spread functions is greater than x'merklich limited in a technically costly manner and thus the harmful halos at least noticeably limited are restricted. In this case, x 'is preferably selected such that x ^' corresponds to n times the pixel size and thus the lateral extent of a pixel of the image recording unit of the camera, where n ≦ 20 and in particular n ≦ 5 is selected. If n is chosen to be smaller, the halos will be so strong. With a choice of n between 10 and 20, a significant reduction in glare due to undesired halos can already be achieved, whereas with a value of n below 5, a very advantageous, particularly marked improvement in the reduction can be achieved.

Within the scope of the invention, it has been found that the halos are not attributable to manufacturing defects or production limits of the objectives, as previously assumed, but are a principal consequence of the acquisition of images with a high contrast range or high dynamic range (HDR images), for example, using CMOS image recorders, ie to go back to physical limits. At the heart of the invention is the use of a special intensity attenuator, e.g. a special aperture or filter to produce a steeply sloping point spreading function. The operation of the invention will be explained with reference to the accompanying figures.

The invention has the advantage that a considerable image improvement is achieved without elaborate digital image processing, thus creating a cost-effective and rapid avoidance or reduction of halos.

The invention will be explained in more detail with reference to an embodiment shown in FIGS. The features described therein may be used alone or in combination with each other to provide preferred embodiments of the invention. Showing:

Fig. 1 is a schematic representation of the point spread function of a diffraction-limited point image in linear scale.

Fig. 2 shows the schematic diagram of a monochrome

Point spreading function of a diffraction limited point image of Fig. 1 in logarithmic scale.

3 shows the basic representation of the point spread function of a diffraction-limited point image in linear scaling of FIG. 1 and a Gaussian function.

Fig. 4 shows the schematic diagram of a polychrome

Point spreading function and a monochrome point spread function of a diffraction-limited point image of FIG. 2 in logarithmic scaling and the Gaussian function of Fig. 3rd

5 shows a first exemplary embodiment of an intensity attenuator according to the invention.

6 shows a second embodiment of an intensity attenuator according to the invention.

At the heart of the invention is the use of a special intensity attenuator, e.g. a special aperture or filter to produce a steeply sloping point spreading function. The operation of the invention will be explained with reference to the accompanying figures as follows.

It is known that the resolving power of optical instruments and devices, such as lenses, is given by the diffraction limited point image. In practice, the so-called point spread function (PSF) is additionally widened by aberrations compared to the diffraction-limited case. But she can initially used as a physical boundary for closer consideration of the halo problem.

The point spread function 1 describes in the optics and image processing the effect of band-limiting influencing factors such as diffraction phenomena on diaphragms, aberrations and the influence of the sensor surface or aperture. It indicates how an idealized point-shaped object would be imaged by an optical system. For microscopes, the width of the point spread function limits the achievable resolution.

The point spread function 1 thus describes the response of an imaging system to a point light source or a point object. Another common expression for the point spread function is system impulse response function. The point spreading function can be considered in many contexts to be an extended blob representing an unresolved object. Functionally expressed, the point spread function is the modulation transfer function in the spatial domain. It is a useful concept in the fields of Fourier optics, imaging in astronomy, electron microscopy, and other imaging techniques such as 3D microscopy (for example, in convocal laser scanning microscopy) and fluorescence microscopy. The extent of expansion or smearing of the dot object in the image is a measure of the quality of the imaging system. For incoherent imaging systems, the imaging process is linear and is described by a linear system theory.

As a result of the linear property, the image of an object can be calculated by decomposing the object into parts, mapping each part and the results be summed up. When the object is divided into discrete point objects of different intensities, the image is calculated as the sum of the point spread functions of all points. Since the point spread function is typically given entirely by the imaging system, the overall image can be described by the known optical properties of the system. This process is usually described by a convolution function. In image processing, for example in microscopy and astronomy, the knowledge of the pulse-spreading function of the imaging system is very important in order to recover the original image by unfolding it.

The Airy function 1 describes the appearance of a star, a point light source, when viewed with a telescope for a single wavelength (monochromatic light). The ideal point image becomes a series of concentric waves forming the halo due to the limited aperture and wave nature of the light.

It is known that the diffraction-limited point spread function 1 for a circular aperture lens has the profile y (r) = (2 x Jl (r) / r) x (2 x Jl (r) / r), where Jl (r) denotes the Bessel function of the first order. This process is also referred to as Airy function 1 and is shown in FIG. 1 for a single wavelength. FIG. 1 shows the (normalized) intensity I as a function of the lateral coordinate x. The function has a main maximum 2 at x = 0, followed by x secondary maxima as the lateral coordinate increases. In the linear representation in FIG. 1, only the first secondary maximum 3 and, as an indication, the second secondary maximum 4 can be recognized. This classic function is reproduced in optics textbooks as in Fig. 1, which is entirely appropriate for the up to a few years ago used recording media and image sensor with reasonably linear characteristics.

In the case of a CMOS imager or a CMOS camera, however, a logarithmic representation of the point spread function 1 is more appropriate. This is shown in Fig. 2. Here you can see that the Airy feature 1 seems to be much wider for a single wavelength. One recognizes the secondary maxima 3, 4, 5, 6, 7 etc. more clearly. In practice, this means that the point image of a very intense source is still very far from its actual location, i. is recognized by the image sensor or by a pixel of an image sensor at large lateral coordinates x, because the intensity I contained in these secondary maxima 3, 4, 5, 6, 7 etc. is also shown in the image. With a dynamic range of the order of 100 dB, the image size of a point object, depending on the aperture ratio of the camera and the pixel size, can easily be 50-100 pixels. The point image is thus outshined at the outer edge by the secondary maxima 3, 4, 5, 6, 7, etc., which also appear in the HDR image, it forms the HaIo. Aberrations broaden the point image even further.

It follows that the halo phenomena can not be eliminated even with a perfectly corrected and manufactured objective, since they are based on mechanisms of image acquisition of the image with high contrast values, ie the CMOS image recording, by which in the secondary maxima 3, 4, 5, 6, 7, etc., which are not interfering with image representation with a linear characteristic intensities I are shown in the image. The maxima and minima shown by way of example occur only when using a single wavelength (monochromatic light). In non-monochromatic images, they are smeared or together form the envelope of the individual different wavelengths used, which is called a polychrome point spread function.

A possible solution would be the above-mentioned downstream digital image processing, but would require a considerable amount of processing power and thus ECU costs and image processing time.

The solution according to the invention is essentially based on the fact that the diffraction-limited, incoherent point spread function can be calculated from the Fourier transformation of the pupil function (the aperture). Furthermore, it is known from astronomy that an annular pupil function, as e.g. for reflecting telescopes, the central area of the Airy function is narrowed. In scientific literature this is called apodization. Unfortunately, the point spread function of a mirror telescope falls even flatter.

In the context of the invention, therefore, a pupil function had to be found whose Fourier transform decreases significantly faster compared to the Airy function 1, at least for high lateral coordinates x, which is easy to implement and which leads to only low light losses. The central area may well be slightly wider than the diffraction-limited point spread function 1; as long as it is not wider than a typical pixel this is not noticeable. Such a pupil function is eg the Gaussian function 8 with the curve e ^{~ aχ2} . It has the pleasant property that its Fourier transform is again a Gaussian function 8. Overall, this falls much steeper than the Airy function 1 and, at a certain lateral value χ ^{v, is} smaller than the polychrome point spread function from the different monochrome point spread functions, as can be seen in FIGS. 3 and 4. However, this leads to an undesirable broadening of the central area. The exact parameters are of subordinate importance, it is important that the function selected in practice decreases in principle more than the polychrome point spread function and / or thus from a lateral value x 'resulting from the given pupil function point spread function (PSF) smaller in particular significantly smaller is the polychrome point spread function without intensity attenuator.

It is clear from FIG. 4 that it is not necessary for the pupil function or its Fourier transform for all lateral coordinates to be greater than a predetermined lateral value x 'to be smaller or substantially smaller than the polychrome point spread function 10 from a plurality of monochrome point spread functions 1 , Halo ringing is achieved as soon as typically one or more of the sub-maxima 3, 4, 5, 6, 7, etc. of the monochrome point spread functions 1 are suppressed and thus the values of the pupil function x are smaller or substantially smaller than the polychrome point spread function for the corresponding lateral coordinates x 10 is.

As explained above, it can be harmless if the central area of the pupil function or its Fourier transform is slightly wider than the diffraction limited polychrome point spreading function 10.

A preferred pupil function or its Fourier transform is the Gaussian function 8. Other advantageous embodiments can consist in that the pupil function or its Fourier transform is one of the following functions or a combination of the following functions: rectangular, triangular or Bartlett, Blackman, Connes, Cosinus, Hamming, Hanning, Kaiser, Nuttall, Welch, power, logarithm or exponential function, preferably a Gaussian function. The mentioned functions are known from the literature for the application of apodization in digital signal processing,

An intensity attenuator according to the invention can be designed, for example, as a diaphragm or as a filter 9. In this case, the filter 9, as well as the aperture, be a separate filter, which is set in front of, behind or in the lens, but it can also be formed as a coating, in particular as vapor deposition of a lens of the lens.

A Gaussian or other pupil function can be realized, for example, by the corresponding evaporation of an objective lens. If this were carried out correctly, as shown for example in FIG. 5, where the light attenuation of the filter 9 has a course in which the light attenuation increases from the optical axis to larger lateral coordinates x, then 50% light loss is to be expected. In practice, it is sufficient to apply an annular filter with a gradual profile only at the edge region, as shown in FIG. 6, where the filter 9 has an extended central region 10 in which the light is not attenuated. The filter 9 shown there has an annular course of the Light attenuation on. The width of this ring 11 results from a consideration of the remaining light intensity of the lens and the size to which the halos are to be limited and the tolerable width of the main maximum 2. In general, a filter according to the invention has a more or less extended central region 10, in the Transmissivity is high, and to the peripheral areas, ie to larger lateral coordinates, a lower transmissivity.

An objective according to the invention can advantageously be used with a camera, with an objective and with an image recorder for recording the image imaged by the objective. The camera is preferably an HDR camera designed as a CMOS camera with a high light intensity dynamic range and / or a high resolvable contrast ratio. The imager may advantageously be a CMOS imager which has a high light intensity dynamic range and / or is designed to capture an image with a high contrast ratio. For example, the light intensity dynamic range of the camera and / or the image sensor may be more than 60 dB, preferably more than 75 dB, and more preferably more than 90 dB. In other applications, the contrast ratio that can be resolved by the camera and / or the image recorder can advantageously be more than 20,000: 1, preferably more than 50,000: 1, and preferably more than 80,000: 1.

The image recorder used can in principle have any, even a linear sensitivity. Preferred embodiments are those in which the image recorder and / or the image enhancement prefer a non-linear sensitivity has a logarithmic sensitivity. A particularly preferred image sensor is a CMOS sensor.

A preferred field of application of the invention is night-vision systems, in particular infrared-sensitive night-vision systems in motor vehicles. In this case, the night vision system can be equipped with or without illumination, for example, an infrared light illumination for illuminating the scene to be recorded.

LIST OF REFERENCE NUMBERS

1 point spread function or Airy function

2 main maximum

3 First secondary maximum

4 Second secondary maximum

5 Third secondary maximum

6 Fourth secondary maximum

7 fifth secondary maximum

8 Gaussian function

9 filters

10 Central area

11 ring

I intensity x lateral coordinate

Claims

claims

1. A camera lens, in particular for a night vision system in a motor vehicle with a camera, which as a function of the lateral coordinate (x), relative to a direction perpendicular to the optical axis of the lens, a diffraction-limited point spread function (PSF), (1) characterized in that the lens comprises an optical intensity attenuator arranged in the imaging beam path, which attenuates the light intensity (I) imaged in the beam path of the objective in a lateral direction according to a predetermined pupil function, so that the resultant point spread function (PSF) increases with increasing lateral intensity Coordinate (x) is smaller than the diffraction-limited polychromatic point spread function (1) of the objective without the intensity attenuator from a certain value x '.

2. Camera lens according to claim 1, characterized in that x 'corresponds to n times the pixel size of an image recording unit of the camera and n ≤ 20, in particular n ≤ 5.

3. Camera lens according to one of the preceding claims, characterized in that the pupil function or its Fourier transform is one of the following functions or a combination of the following functions: Rectangle, Triangle or Bartlett, Blackman, Connes, Cosinus, Hamming, Hanning, Kaiser, Nuttall, Welch, power, logarithm or exponential function, preferably a Gaussian function (8).

4. Camera lens according to one of the preceding claims, characterized in that the intensity attenuator is designed as a filter (9) or aperture.

5. Camera lens according to the preceding claim, characterized in that the filter (9) is a separate filter, which is placed in front, behind or in the lens.

6. Camera lens according to claim 4, characterized in that the aperture is a separate aperture, which is set before, behind or in the lens.

7. Camera lens according to one of claims 4 to 6, characterized in that the filter (9) or the diaphragm is formed as a coating, in particular as vapor deposition of a lens of the lens.

8. Camera lens according to one of claims 4 to 7, characterized in that the filter (9) or the diaphragm has an annular profile of the light attenuation.

9. Camera lens according to one of claims 4 to 8, characterized in that the light attenuation of the filter (9) or the diaphragm has a course in which the light attenuation from the optical axis to larger lateral coordinates (x) increases towards.

10. Camera lens according to one of claims 4 to 9, characterized in that the light attenuation of the filter (9) or the diaphragm has a gradual course, in which the light attenuation of the optical axis to larger lateral coordinates (x) increases towards.

11. Camera lens according to one of claims 4 to 10, characterized in that the filter (9) or the diaphragm has a central region (10) in which the light is not attenuated.

12. A camera with a lens and with an image sensor for receiving the image imaged by the lens, characterized in that it comprises a lens according to one of the preceding claims.

A camera according to the preceding claim, characterized in that the values of the pupil function or its Fourier transform for those lateral coordinates (x) which correspond to the extent of a typical pixel of the image recorder are greater than the values of the point spreading function (1) for them lateral coordinates (x).

14. A camera according to any one of claims 12 to 13, characterized in that it is an HDR camera with a high light intensity dynamic range and / or a high resolvable contrast ratio.

A camera according to the preceding claim, characterized in that the imager is an HDR imager which has a high light intensity dynamic range and / or is adapted to capture an image having a high contrast ratio.

16. A camera according to any one of claims 12 to 15, characterized in that the light intensity dynamic range of the camera and / or the image sensor is more than 60 dB, preferably more than 75 dB and more preferably more than 90 dB.

17. A camera according to any one of claims 12 to 16, characterized in that the resolvable by the camera and / or the image sensor contrast ratio more than

20,000: 1, preferably more than 50,000: 1 and preferably more than 80,000: 1.

18. A camera according to any one of claims 12 to 17, characterized in that the image sensor and / or the image enhancement has a non-linear sensitivity, preferably a logarithmic sensitivity.

19. A camera according to any one of claims 12 to 18, characterized in that the image sensor is a CMOS sensor.

20. A camera according to any one of claims 12 to 19, characterized in that it is an infrared-sensitive camera of a night vision system.

21. Camera according to the preceding claim, characterized in that the night vision system comprises a lighting, preferably an infrared light illumination, for illuminating the scene to be recorded.

22. A camera according to any one of claims 20 to 21, characterized in that it is a camera of a night vision system in a motor vehicle.