US 20060091813 A1
An automatic vehicle headlamp dimming system which includes an optical system and an imaging processing system. The optical system is configured to discriminate between headlamps and tail lamps, and focus the light rays from the headlamps and tail lamps on different portions of a pixel sensor array. The optical system as well as the image processing system provides for relatively increased discrimination of headlamps and tail lamps of other vehicles and also enables the high beam headlamps of the control vehicle to be controlled as a function of the distance as well as horizontal angular position of other vehicles relative to the controlled vehicle.
1. A control system for automatically controlling the state of the head lamps of a controlled vehicle, the control system comprising:
an optical system for imaging external sources of light within a predetermined field of view, the optical system including an image array sensor and two or more lenses configured to image said predetermined field of view onto two or more corresponding portions of said array; and
an image processing system for processing images from said optical system and providing a control signal for controlling the head lamps as a function of the relative output of the pixels imaging said external sources of light.
2. The control system as recited in
3. The control system as recited in
4. The control system as recited in
5. The control system as recited in
This application is a continuation of U.S. patent application Ser. No. 11/180,346, entitled “CONTROL SYSTEM TO AUTOMATICALLY DIM VEHICLE HEADLAMPS” filed on Jul. 13, 2005, by Joseph S. Stam et al., which is a continuation of U.S. patent application Ser. No. 10/677,453, filed on Oct. 2, 2003, now U.S. Pat. No. 6,919,548, which is a continuation of U.S. patent application Ser. No. 09/874,197, filed on Jun. 5, 2001, now U.S. Pat. No. 6,653,614, which is a continuation of U.S. patent application Ser. No. 09/151,487, filed on Sep. 11, 1998, now U.S. Pat. No. 6,255,639, which is a continuation of U.S. patent application Ser. No. 08/831,232, filed on Apr. 2, 1997, now U.S. Pat. No. 5,837,994. Priority under 35 U.S.C. § 120 is hereby claimed as to each of the above applications and the entire disclosures of each are incorporated herein by reference.
The present invention relates to a system for automatically dimming vehicle high beam headlamps.
Regulations set forth by the United States Department of Transportation (DOT) regulate the light emissions of vehicle high beam headlamps. Various state regulations are used to control the amount of glare experienced by drivers of other vehicles whether the vehicle is traveling in the same direction as the controlled vehicle or in an opposite direction.
Known vehicle high beam headlamp emissions in accordance with the DOT regulations provide an intensity of 40,000 cd at 0 degrees, 10,000 cd at 3 degrees, 3250 cd at 6 degrees, 1500 cd at 9 degrees, and 750 cd at 12 degrees. An example of such an emission pattern is illustrated in
Various known head light dimmer control systems are known in the art. In order to prevent the drivers of other vehicles from being subjected to excessive glare levels, such automatic headlamp dimmer systems must sense both the head lights as well as the tail lights of other vehicles. While many known systems are adequately able to detect headlamps of oncoming vehicles, such systems are known to inadequately sense tail lights of vehicles traveling ahead of the control vehicle. As such, such systems are not able to automatically dim the high beam headlamps in time to prevent drivers of the vehicles travelling in the same direction as the controlled vehicle being subjected to excessive glare levels.
U.S. Pat. No. 5,537,003, assigned to the same assignee of the present invention, discloses an automatic headlamp dimming system which includes an optical system for sensing tail lamps as well as headlamps. The '003 patent discloses a single photo diode with a mechanical scanning arrangement for scanning a predetermined field of view. Although the system provides relatively suitable sensing of headlamps as well as tail lamps, the optical subsystem is rather complicated and expensive to manufacture.
It is an object of the present invention to solve various problems in the prior art.
It is yet another object of the present invention to provide a vehicle headlamp dimming system which eliminates the need for mechanical optical scanning systems.
It is yet another object of the present invention to provide a headlamp dimming system that is adapted to dim the high beam head lights at different distances as a function of the horizontal angular position of another vehicle relative to the controlled vehicle.
Briefly, the present invention relates to an automatic vehicle headlamp dimming system. The system includes an optical system and an imaging processing system. The optical system is configured to discriminate between headlamps and tail lamps and focus the light rays from the headlamps and tail lamps on different portions of a pixel sensor array. The optical system, as well as the image processing system, provides for relatively increased discrimination of headlamps and tail lamps of other vehicles and also enables the high beam headlamps of the control vehicle to be controlled as a function of the distance as well as the horizontal angular position of other vehicles relative to the controlled vehicle.
These and other features, advantages, and objects of the present invention will be further understood and appreciated by those skilled in the art by reference to the following specification, claims, and appended drawings.
These and other objects of the present invention will be readily understood with reference to the following specification and attached drawing, wherein:
The automatic headlamp dimming system in accordance with the present invention includes an optical system as illustrated in
The lens 103 may be formed with a red filter dye for transmitting light with wavelengths greater than 600 nm and focusing red light rays 101 from tail lamps onto one-half of the image array sensor 106. The red filter dye causes the lens 103 to absorb all light rays at the blue end of the visible spectrum and transmit light rays at the red end of the spectrum. As such, the amount of light transmitted from non-red light sources, such as headlamps, is greatly reduced while light rays from tail lamps are fully transmitted through the lens 103. As such, the relative brightness of the light rays from tail lamps imaged onto the image array sensor 106 is greatly increased.
The lens 104 may be formed with a cyan-filtered dye for transmitting light with wavelengths less than 600 nm. The lens 104 is used to focus the light rays onto the other half of the image array sensor 106. The cyan filter dye has a complementary effect to the red filter described above. In particular, the red filter dye absorbs light from the blue end of the visible spectrum while transmitting light from the red end of the spectrum. As such, most of the light from sources, such as head lights, is transmitted through the lens 104 while virtually all of the light emanating from tail lamps is blocked.
Both headlamps and tail lamps emit a substantial amount of infrared light. By utilizing lenses with a filter dye or separate filters which inhibit light at wavelengths greater about 750 nm, the infrared light transmitted by the headlamps and tail lamps will be substantially blocked by the lenses 103 and 104. By eliminating infrared light, the ratio between intensity between red lights imaged through the red filter and red light imaged through the cyan filter will be substantially increased.
The use of the red and cyan dyes for the lenses 103 and 104 is merely exemplary. The filter characteristics of the lenses 103 and 104 are selected to optimize the sensitivity of the device to specific light sources. For example, the headlamps or tail lamps in new vehicles may be replaced with alternative light sources with different spectral composition, for example, with high intensity discharge headlamps and light emitting diode tail lamps requiring different filter characteristics. Depending on the spectral characteristics of the headlamps and tail lamps, transparent lenses 103 and 104 may be utilized with separate color filters.
The lenses 103 and 104 may be formed as acrylic spherical lenses. Alternatively, the lenses 103 and 104 may be formed as aspherical lenses in order to minimize color dispersion and spherical aberration present with spherical lenses. Complex lenses formed from both spherical and aspherical lenses are also contemplated.
A single lens may also be used in place of the separate lenses 103 and 104. The use of a single lens may be used to image the field of view onto a full or partial color image array sensor containing pigmentation on the individual pixels in the array.
As shown best in
As mentioned above, the light rays 101 transmitted through the lens 103 are imaged onto one-half of the image array sensor 106 while the light rays 102 transmitted through the lens 104 are imaged onto the other half of the image array sensor 106. In order to provide such spatial segregation of the light rays transmitted through the lenses 103 and 104, the lens holder 105 is provided with cutouts 107 and preferably formed or coated with a light absorbing material. These cutouts 107 prevent light rays beyond the desired maximum vertical angle transmitted through the red lens 103 from being imaged onto the portion of the image array sensor 106 reserved for the light rays 102. Conversely, the cutouts 107 also prevent light rays transmitted through the lens 104 from being imaged onto the portion of the image array sensor 106 reserved for the light rays 101.
The field of view of the optical system is defined by the configuration of the lenses 103 and 104 and the cutouts 107 relative to the image array sensor 106. For example, an exemplary field of view of 10 degrees in the vertical direction and 20 degrees in the horizontal directions may be created by the configuration set forth below. In particular, for such a field of view, the lenses 103 and 104 are selected with a diameter of 1.5 mm with a small portion cut away to allow the lenses 103, 104 to be positioned such that their centers are separated by 1.0 mm. The lens 103 is positioned 4.15 mm from the image array sensor 106 while the lens 104 is positioned 4.05 mm away. Both the front and rear surface radii of the lenses 103 and 104 are 4.3 millimeters with a 0.2 millimeter thickness.
As best shown in
The configuration described above is thus able to baffle light outside of the desired horizontal and vertical field of view. In particular,
The image array sensor 106 may be CMOS active pixel image sensor array, for example, as disclosed in U.S. Pat. No. 5,471,515, hereby incorporated by reference and available from Photobit LLC of La Crasenta, Calif. CMOS active pixel image sensors provide relatively high sensitivity and low power consumption as well as the ability to integrate other CMOS electronics on the same chip. The image array sensor 106 may be a 50×50 40 μm pixel array. The number of pixels in the image array sensor 106 is selected such that not all pixels fall within the area that the lenses 103 and 104 project onto. The extra pixels allow for simple correction for mechanical misalignment by offsetting the expected image location.
The image array sensor 106 provides spatial information regarding light sources within the field of view. The number of pixels present in the array is selected to obtain sufficient spatial detail although the size of the array is not limited and may be selected, and may even be dictated by physical and economic limitations.
The image array sensor 106 must be sensitive to accurately detect tail lights at several hundred feet. Such sensitivity may be achieved by lengthening the amount of time the photosites in the array are exposed to light during a frame period. A frame period is selected to enable the array to capture and transfer a frame to the image processing system in a short enough time to allow the image processing system to detect another vehicle entering the field of view. A short time period also limits the amount of motion within a frame during the integration period and thus produces a relatively more accurate instantaneous image.
The use of a pixel array also provides other benefits. For example, as mentioned above, the light integration time to capture a frame can be varied. Such a feature allows the system to provide optimal results in varying degrees in darkness. Another important aspect of an image array sensor is the ability to utilize subsets of the pixels within the array or an individual pixel. As such, as the window size is decreased, the readout rates can be increased. Such a feature allows the system to discriminate ambient light sources, such as street lamps. In particular, such a feature allows the system to locate a light source within the frame and capture several samples of the light sources at a rate several times greater than 60 Hz. In particular, if the samples exhibit 120 Hz intensity modulation, the light source is likely a street lamp or other light source powered from a 60 Hz AC power supply. If the light source is not modulated, the light source is likely powered by the vehicle's DC power supply.
Another potential benefit of the image array sensor is that it allows the field of view immediately in front of the vehicle to be imaged by a higher pixel resolution. Thus, the system may be configured such that the effective pixel resolution decreases as the angle of the vehicle relative to the control vehicle increases thus reducing the amount of processing time in those areas. Such a configuration reduces the sensitivity of the device to light sources from reflective stationary objects on the side of the road.
An image array sensor could be manufactured in which the pixel pitch is varied as a function of the area in the field of view that the pixel images. For example, pixels imaging the space corresponding to horizontal angles within 3 degrees of the center of the vehicle may be provided with a 10 μm pixel pitch. Pixels imaging horizontal angles between 3 and 6 degrees may be provided with a 20 μm pixel pitch, while those imaging angles greater than 60 degrees may be provided with a 40 μm pitch. While such a configuration may not increase the sensing area, the ability to resolve detail increases; an important aspect in considering the relative size of a tail lamp at a relatively large distance. For example, a 4 1/2 inch diameter tail light at a distance of 200 feet subtends an angle of less than 0.11 degrees. If a 50×50 image array sensor is used to image a 20 degree field of view, the tail lamp would subtend approximately 8% of the total area imaged by the pixel.
A tail lamp is relatively brighter than its ambient surroundings; however, the red light contributed by the tail light is diluted by the ambient light at such a distance. Such a factor is critical when comparing the amount of red light in a given area to the amount of non-red light in the same area. When the area of space compared is large relative to the light source, the percentage of red light is diminished. By comparison, if 10 μm pixels are used in the center of the array 106 instead of 40 μm pixels, the tail lamp would subtend 0.04% of the total areas, an improvement of 16 times.
Image Processing System
The image processing system is illustrated in
The flow chart for the headlamp control is illustrated in
At the beginning of each cycle, two images are acquired through the lenses 103 and 104. In step 302, the images from the lenses 103 and 104 are analyzed to detect tail lamps. Another image is acquired in step 303 through the lens 104. The image acquired through the lens 104 is acquired with a low enough gain to detect oncoming head lights while rejecting lower light level reflections and nuisance light sources. After the images are analyzed, the system checks for very bright lights in the image indicating the sudden appearance of vehicle headlamps or tail lamps within the field of view, as is the case when a car turns in front of the controlled vehicle in step 305. If bright lights are sensed, the device dims the headlamps 206 immediately and bypasses the time verification as discussed below. The cycle is then repeated. If there were no bright lights, the system proceeds to step 307 to determine if there are any headlamps or tail lamps in the image.
In order to confirm the presence or lack of presence of a headlamp or tail lamp in a frame, an undim counter and a dim counter are used. These counters provide verification of a particular light source whether from a tail lamp or headlamp from consecutive frames before signaling the headlamp control unit 205 to dim or undim the headlamps 206, except as described above, when a bright light is detected. By providing verification, anomalies within the device or in the image will not cause spurious operation of the headlamps 206.
The dim counter is incremented each time a frame with a headlamp or tail lamp is detected until the number of required consecutive frames to take action are reached. The dim counter is set to 0 each time a clear frame is processed. The undim counter is incremented with each clear frame and set to 0 with each frame containing a headlamp or tail lamp. The actual number of consecutive frames required to dim or undim is determined by the overall speed of the device. The more frames used for verification, the less susceptible the system will be to noise and anomalies. However, the device must be able to react quickly to be effective so the number of verification frames is kept relatively low. Whenever a headlamp or tail lamp is detected in step 307, the undim counter is set to 0 in step 308. In step 309, the system checks whether the headlamp 206 high beams are on. If the high beams are off, no further action is required and the cycle is repeated as indicated by step 317. If the high beams are on, the dim counter is incremented in step 310. After the dim counter is incremented in step 310, the system checks in step 311 if the dim counter has reached the number of consecutive frames required to dim the headlamps 206. If so, the system proceeds to step 306 and dims the headlamps 206 and resets both the dim and undim counters and repeats the cycle. Otherwise, the system repeats the cycle and proceeds to step to 317.
In step 307, if there are no headlamps or tail lamps in the image, the dim counter is set to 0 in step 312. Subsequently, in step 313, the system determines whether the high beams 206 are on. If the high beams are on, the system exits repeats the cycle in step 317. In step 313 if the brights are not on, the undim counter is incremented. After the undim counter is incremented, the system checks in step 315 whether the undim counter has reached the number of consecutive clear frames required to activate the high beams 206. If so, the high beams are turned on in step 316, and the cycle is repeated. If the undim counter is less than the required number for activating the bright headlamps 206, the system repeats the cycle in step 317.
The flow diagram for tail light processing is illustrated in
The threshold value may be computed in a variety of ways. For example, it can be a predetermined fixed number or a number that is a function of the current image sensor and ADC settings. The threshold value can also be determined by computing a threshold as a factor of the average pixel intensity of the entire image which would help eliminate variances caused by changing ambient light sources. In addition, the pixel value may be compared to the average of the pixels in the immediate area of the pikel of interest. This local average method prevents relatively large, moderately bright spots in the image from being seen as vehicle light sources. More particularly, distant tail lamps subtend less than one pixel and thus will only have moderate brightness. Large spots in the image with moderate brightness are most likely caused by reflections from large objects. Close tail lamps which subtend many pixels will have a saturated center which will be brighter than the surrounding pixels allowing the same method to detect them as well.
The threshold may also be determined by varying the threshold spatially by way of a lookup table or computation. However, the threshold should be determined so that dimming occurs appropriately for the dimmest tail lights allowed by the DOT standards. Distant vehicles are subjected to the most intense portion of the controlled vehicle high beam, thus requiring dimming only directly in front of the controlled vehicle as indicated in
A similar approach can be taken for varying the threshold for pixels in imaging areas of space and angles above and below the center. However, a more conservative approach can be taken when determining the tail light sensitivity relative to the vertical angle since vehicles tend to move more frequently and rapidly in vertical directions due to hills and bumps in the road. Therefore, by specifying relatively tight vertical thresholds may cause the bright headlamps 206 to switch on and off as the vehicle moves several degrees up and down.
A hysteresis multiplier may be applied to the threshold to prevent oscillations of the headlamps 206 when the light source has a gray scale value at or near the threshold. Thus, if the bright headlamps 206 are off, the threshold will be lower for all pixels to prevent the bright headlamps from coming back on, even if the faintest tail lamps are present in the image. However, if the bright headlamps 206 are on, the threshold should be higher so that only tail lamps of sufficient brightness are sensed to indicate that the car is within the dimming range to cause the headlamps 206 to dim.
One of the biggest problems facing the detection of the tail lamps is the nuisance red light reflected from corner cube reflectors commonly found as markers on the side of the road and on mailboxes. The variable threshold method mentioned above helps eliminate some of this noise. However, when a vehicle approaches a reflector at the proper angles, it is relatively impossible to distinguish a red reflector from a tail lamp. Fortunately, by examining successive frames and investigating the motion of these objects over time, such reflections can be filtered. By storing the location of the tail lamps and images over time or by sensing small regions of interest where the tail lamp is located several consecutive times, the device can look for rightward motion and determine if the light source is a reflector. Additionally, the speed at which the control vehicle overtakes a stationary object is much greater than the relative rate a vehicle would overtake another moving vehicle. Thus, the rate of increase in brightness of the object would be typically much greater for a stationary reflector than for another vehicle. This rate of change in brightness coupled with rightward horizontal motion can be used as signatures to reduce the number of false tail lamps detected.
A computationally simpler method of analyzing spatial motion of a light source is to simply take several consecutive regions of the local region of interest where the light source is located. Motion in the vertical and horizontal directions is relatively slow for tail lamps of a leading vehicle. Simply sampling a pixel a few consecutive times to see if the tail lamp is present in all samples can adequately eliminate objects which rapidly move within the image.
The flow diagram for tail lamp processing is illustrated in
As discussed above, the other criteria for tail lamp detection relates to the relative brightness of the pixel relative to neighboring pixels. Thus, in step 322, the system calculates the average gray scale value of neighboring pixels. If it is determined in step 323 that the pixel gray scale value for the pixel imaged through the lens 103 is n % greater than the average gray scale value of the neighboring pixels, the system proceeds to step 324 and adds the pixel to the tail lamp list for future frames of reference. Otherwise, the system moves to step 319 and moves the next pixel. In steps 325 and 326, the systems determines whether or not the red light detected is a tail lamp or a reflector, as discussed above. If it is determined that the light is a reflector, the system proceeds to step 327 and moves on to the next pixel. Otherwise, the headlamps are dimmed in step 328.
The flow diagram for head light processing is illustrated in
An additional concern with headlamp detection arises from the rapid decrease in distance between oncoming vehicles which becomes especially critical when an oncoming vehicle suddenly enters the controlled vehicle's field of view, for example, when turning a corner or in a similar situation. For this reason, an additional flag is used to cause the vehicle to immediately dim the bright headlamps and bypass any verification if the light source is above certain absolute high-level brightness thresholds.
The primary nuisance light source complicating headlamp detection comes from overhead lights, such as street lights and electrically illuminated street signs. One method of eliminating such nuisance light sources is to analyze their motion. In particular, all overhead street lamps will move vertically upwards in the image as the controlled vehicle is moving. Analyzing this motion provides an efficient method of detecting some street lamps. Unfortunately, distant street lamps appear to be at almost the same elevational angles as distant head lights and the rate of vertical climb in the image does not become great until the street lamp is closer. However, as discussed above, street lighting is AC controlled and thus is subject to 120 Hz intensity modulation. Headlamps powered by DC source do not exhibit this characteristic. Thus, the image array sensor 106 is able to utilize a small number of pixels for taking several rapid consecutive readings in a window. If the window is small enough, the window can read several hundred frames per second. Once the light source is identified in the image, several frames are read out at a rate of 240 Hz or higher. These readings are then analyzed to detect the intensity modulation. If modulation is present, the light source originates from an AC source and can be ignored. Alternatively, a photodiode can be used in conjunction with a low pass filter to determine the ratio of light in the image that was AC modulated to the unmodulated light. If a significant portion of the light source is AC modulated, the light source present in the image is assumed to be from AC light. Otherwise, the light source is assumed to be from a DC source.
The flow diagram for headlamp processing is illustrated in
As discussed above, the system examines light sources as discussed above in steps 338 and 339 to determine if the light source is a street lamp. If the system determines that the light source is not a street lamp, the system proceeds to step 340 and sets a flag to cause dimming of the headlamps 206. If the system determines that the light source is a street lamp, the system proceeds to step 341 and moves on to the next pixel.
Traditional vehicle lamps systems have the option of the bright lamps being either on or off. The present invention is readily adaptable for use with a headlamp system where the brights can be activated to vary the brightness based on the distance of other vehicles in the field of view. In such an embodiment, the brightness of the headlamps may be varied by various techniques including modulating the duty cycle of the headlamp in order to reduce or increase the overall brightness level.
Variable intensity headlamps also result in better noise filtration. In particular, whenever a light source is detected which causes the brightness of the controlled headlamps of the vehicles to be decreased, other images can be detected to determine if the intensity of these other light sources decreases by a similar amount. If so, the system would be able to determine that the light source is a reflection from the vehicle's headlamps. Such information can be used as feedback to provide a relatively efficient means for eliminating nuisance light caused by reflections of the control vehicle headlamps. In such an embodiment, the brightness threshold discussed above would not be used. More particularly, the brightness of the brightest headlamp and tail lamp in the images is used to determine the brightness of the controlled vehicle's headlamps. The brighter the headlamps or tail lamp in the images, the dimmer the controlled headlamps.
Obviously, many modifications and variations of the present invention are possible in light of the above teachings. Thus, it is to be understood that, within the scope of the appended claims, the invention may be practiced otherwise than as specifically described above.
The above description is considered that of the preferred embodiments only. Modifications of the invention will occur to those skilled in the art and to those who make or use the invention. Therefore, it is understood that the embodiments shown in the drawings and described above are merely for illustrative purposes and not intended to limit the scope of the invention, which is defined by the following claims as interpreted according to the principles of patent law, including the doctrine of equivalents.