Brightness control on video cameras
Patent Information
- Application Number
- MYPI1998000206
- Authority / Receiving Office
- MY · MY
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 1997-01-17
- Filing Date
- 1998-01-17
- Publication Date
- 1998-07-17
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for regulating exposure in stationarily mounted video cameras, such as those used in automatic monitoring systems, often result in incorrect exposures under difficult lighting conditions, as they rely on light sensors outside the camera's viewing area, failing to accurately adapt to changing conditions like a wet, glistening road.
The method calculates a correction factor for exposure parameters directly from the image excerpt observed by the camera, using grey-scale value histograms and cross-correlations across selected measurement zones, ensuring that exposure adjustments are made based on the actual light-sensitive plane, and includes checks for plausibility and overdriving criteria to prevent errors from moving objects.
This approach allows for optimal exposure control even in challenging lighting conditions, effectively compensating for day-night differences and preventing exposure errors, ensuring continuous and accurate adaptation of camera settings in applications like automatic traffic toll collection systems.
Abstract
Description
A method of controlling the exposure of stationarily mounted video 5 cameras (40) is disclosed rendering it possible to obtain optimal exposure also in difficult lighting situations. To this end, measuring ranges of the image are evaluated and based on the said evaluation a correction factor influencing the exposure parameters of the camera is calculated.Figure 4Brightness control on video camerasPrior ArtThe invention is based on a Method for regulating exposure parameters of stationarily mounted video cameras, e.g. for travel monitoring, according to the preamble of the main claim.The British Patent GB 21 86 461 has already disclosed a method for regulating the exposure for a stationary video camera. This involves cameras which are used in automatic monitoring systems. These camera systems monitor an object throughout the 24 hour period. This means that the exposure conditions for the area to be observed can, in general, also change drastically over the course of the day. Automatic monitoring of the exposure must be built into the camera systems. The British patent specification proposes a sensor which detects the light conditions through the camera lens. Using the measured values, the lighting parameters of the camera are set and tracked. The sensor receives its exposure data from an area which is not the viewing area of the video camera. This has the disadvantage that, under very difficult light conditions prevailing in the area to be observed, incorrect exposures can occur. Thus, it ia possible, for example, that a glistening road wet with rain will require completely different exposure preconditions from those that the sensor can identify from its measurement of an image excerpt, e.g. an area of sky.Automatic gain controls are necessary, which allow continuous adaptation of the exposure parameters with continual evaluation of the image content.A method for detecting the image width in a video signal is already known per se from the International Application WO-A-87 / 03768, in which method the exposure parameters of a video camera are set using a correction factor determined by comparison of grey-scale values of sequences of pixels of the preceding image and of the current image.The invention in accordance with the main claim is based on the object of avoiding the abovementioned disadvantages .Advantages of the InventionBy contrast, the method according to the invention, having the characterizing features of the main claim, has the advantage that the exposure parameters for the video camera are obtained exactly from the image excerpt observed by the camera, and exactly on the light-sensitive plane. Moreover, it is used for applications in which the major differences between day and night is [sicJ advantageously compensated for by a flash unit. As a result, it is possible to obtain a good exposure even under very difficult light conditions, if only individual images are evaluated for the purpose of tracking the exposure parameters.Advantageous developments and improvements of the method specified in the main claim are possible by virtue of the measures presented in the dependent claims.It is particularly advantageous that the method selects measurement zones from the image, and determines a correction factor using grey-scale value histograms and cross-correlations. It is particularly advantageous that the correction factor used for the exposure parameters is calculated as an average value over all the selected measurement zones. Furthermore, it is advantageous that the individual correction factor ia checked with regard to its plausibility. This avoids the situation where e.g. a .vehicle moving through theimage corrupts factor. the exposure and thus the correction Furthermore, it is advantageous that, for monitoring the overdriving of an image, measurement zones arelikewise selected from the image. A number of criteria have to be met for the presence of overdriving. The method sequence according to Claim 6 is advantageously provided.DrawingExemplary embodiments of the invention are illustrated in the drawing and explained in more detail in the description below. Figure 1 diagrammatically shows a grey-scale value histogram, Figure 2 shows a grey-scale value histogram with overdriving, Figure 3 diagrammatically shows the procedure during a cross-correlation, and Figure 4 shows the video camera.Description of the Exemplary Embodiment figure 4 shows a video camera 40 with an objective 46 and a light-sensitive CCD unit 41. The CCD unit is connected to a regulator 42, which is in turn linked to an external controller 43. The external controller 43 is connected to a screen 44. Individual images can be displayed on the screen, from which images measurement tones'45 are selected.The invention's exposure control of a video camera is used e.g. in a system for automatic traffic toll collection. Such a system monitors properly performed accounting operations during passage through a toll station and creates still images of the vehicles. From the still images, e.g. the number· plates are identified by automatic character recognition. The system has to operate round the clock; therefore, the exposure parameters of the camera have to be adapted to the respective light conditions. Exposure parameters aree.g. aperture settings, read-out duration of the CCD cells, etc.' For this purpose, the images from current operation, some of which are recorded using flash light, are evaluated in order to ascertain the present light viewing situation.Firstly, a representative measurement zones (sic] is selected from the electronically recorded still images, which are present as pixel images in the controller 43. A measurement zone 45 is defined as a contiguous sequence of image pixels. In this- case, two different measurement zones are taken as a basis:- measurement zones for determining the correction factor of the exposure, and measurement zones for overdriving monitoring of the camera.The number and the position of the measurement zones is defined during the installation of the system in the controller 43, but can also subsequently be altered interactively at any time. In this case, the measurement zone type can also be altered as desired from a zone for determining the correction factor to a zone for monitoring overdriving. For the measurement zones, once they have been defined, the grey-scale value histogram is determined and stored, e.g. for later determination of the correction factor. Figure 1 shows the frequency 2 of a grey-scale value of the individual pixels of the measurement zone 45 as y-axis plotted against the grey-scale value 1 on the x-axis. For an 8-bit signal, the grey-scale value can assume values between 0 and 255. The spread 3 of the grey-scale value histogram is the standard deviation of the grey-scale values. The expected value is calculated by averaging over the frequency distribution. From such a grey-value histogram, further values such as position4 and value 5 of the maximum of the distribution are calculated and stored.For a first. start-up, the video camera system is initially calibrated. For this, the exposure parameters of the camera, shutter speeds, and also gain, are set in such a way that an ideal exposure of the recording is achieved in conjunction with the current light situation. In the system according to the invention, in addition to visual monitoring of the exposure of the recordings, the grey-scale value histogram of each measurement zone can be immediately displayed. Thus, the result can be monitored as early as during calibration. Once the system is activated, the exposure must be continually monitored. If the exposure of a newly recorded image is intended to be checked, it is assumed during the determination of the correction factor that both the described calibration values and the values of the currently recorded image are present from each measurement zone for the correction factor. Both values are present in the form of grey-scale value histograms. The result of cross-correlation of the grey-scale value histograms is used to compare the exposure of the two images. Seen mathematically, the cross-correlation forms the sum over two comparable functions x and y with a displacement vector x.Figure 3 illustrates the frequency 2 of specific grey-scale values as y-axis against the grey-scale value 1 as x-axis. Two functions x and y are illustrated, both of which have a maximum at the location m. If such functions are used to form the cross-correlation function, then the maximum is situated exactly at x “ 0. That is utilized during the comparison of the exposure of the two images. In this case, it is assumed that the function profile of the cross-correlation does not change greatly given identical image contents but different exposure, rather merely the position of the maximum of the cross-correlation is displaced. It is also assumed of the cross-correlation coefficient, i.e. of the maximum of the cross-correlation, that it remains approximately in the event of altered exposure. The position of the maximum, which is·within the range of greater than or less than zero or is zero, is used as an indicator of the quantitative alteration of the exposure situation. If the image content changed drastically between the two grey-scale value histograms x and y, it would have to be assumed that e.g. a disturbing object is currently crossing the image, and, therefore, the measurement cannot be used for exposure regulation. The correction factor of each measurement zone ie thus the position of the maximum, weighted with the value of the maximum of the cross-correlation. The total correction factor, which alters the exposure, is the negative average value of all the individual correction factors, the said average value being weighted with the areas of the measurement zones. This correction factor specifies the necessary alteration of the camera exposure in grey-scale values. In order to preclude the situation where correction factors in the case of which the measurement is corrupted by objects moving past enter into the weighting, it is necessary for theCorrection factorTotal area of the measurement zones Number of measurement zones of type 1 Area of measurement zone IMaxCorr: Maximum of the cross-correlationMaxPos: Position of the maximum of the cross-correlation correction factor to be subjected to validation. It must be possible to preclude the situation where the image, content has changed within a measurement zone and, therefore, the calculated value is not permitted to be used. The image content of a zona changes e.g. if a vehicle completely or partially fills the zone. This is undesirable for the measurement since the form of the grey-scale value histogram changes drastically as a result. Therefore, a number of criteria are introduced which lead to the rejection of a correction factor of a measurement zone. These criteria are:a) an excessively large difference between expected value and maximum of the cross-correlation of the grey-scale value histograms. This is based on the fact that the form of the cross-correlation does not greatly change given an identical image content. Therefore, maximum of the measurement and expected value of the cross-correlation must be closely adjacent or, in the ideal case, identical. The maximum permitted difference can be set by a threshold in the system presented here.b) a large difference in the standard deviation of stored and current grey-scale value histograms of the measurement zone. The criterion was introduced if the currently recorded image has, owing to a disturbance, a grey-scale value histogram profile that is homogeneous over the entire area. The cross-correlation will then likewise have a homogenous profile, which is why it is possible for expected value and maximum to lie so close together that criterion a) i3 met. In such a case, however, the image contents of the measurement zones are different and the measurement zone must be rejected. Therefore, a comparison of the standard deviations, which represent the spread of the grey-scale value histogram, was carried out. In this case, too, the maximum permitted difference can be set by a threshold.In order to prevent overdriving of the system, measurement zones are selected which serve for monitoring overdriving. The grey-scale value histogram is likewise determined from the measurement zones and the measurement result is used for the following criteria for indication of overdriving:a) the maximum of the grey-scale value histogram lies at the largest possible grey-scale value. Figure 2 shows the frequency of specific grey-scale values 2 against the grey-scale values 1. As illustrated in the solid line in Figure 2, the maximum for such a measurement lies at the maximum grey-scale value. The maximum grey-scale value is ascertained by comparing the largest grey-scale value with the average value of a number of adjacent grey-scale values. A peak is present when the value at the largest grey-scale value is greater than the average value of the adjacent grey-scale values by an adjustable factor.b) in the grey-scale' value histogram, there is a peak at the largest grey-scale value and the expected value of the grey-scale value histogram lies above a certain threshold. This situation is illustrated by the broken line in Figure 2. In this case, there is no absolute maximum at the highest grey-scale value, but rather a generally high level across the entire distribution. If at least one of the criteria a) or b) is met, then the image area of the measurement zone is regarded as overdriven. A plurality of measurement zones are evaluated for monitoring the overdriving. The evaluation is always effected over a plurality of successive images. The images and their grey-scale value histograms are stored irrespective of whether or not they were overdriven. If the current image is identified as overdriven by one or more measurement zones, the following conditions are checked: did the calculation of the correction factor reveal that the image should be made brighter and were an adjustable number of preceding images overdriven. If this condition is met, then alteration of the exposure parameters of the camera is prevented. The entire sequence of exposure regulation is divided into the following steps:1. Definition of size, position and type of the measurement zones in the image area. The system distinguishes between measurement zones for determining the correction factor and those for monitoring overdriving.2. Setting of the optimum exposure.3. Extraction and storage of the reference values of each measurement zone and of the current exposure parameters of the camera (calibration).4. Continuous operation. The regulation of the exposure become [sic] necessary in continuous operation. The sequence of the regulation can be summarized thus:4.1 Selection of the last recorded image or, if the last image is too old, recording of a new image and extraction of the measurement data.4.2 Comparison of the data of the current image with the stored reference data, determination of the correction value from the measurement zones for the correction value.4.3 Checking of the criteria with respect to overdriving from the measurement zone for checking overdriving. If the criteria are met, return to point 1).4.4 Changing of the camera parameters if the magnitude of the correction value exceeds an adjustable value.
Claims
Claims 1. Method for regulating exposure parameters of stationarily mounted video cameras (40), the video cameras recording images which are processed to form electronic pixels, the exposure parameters of the video cameras (40) being set using a correction factor, characterized in that the correction factor is determined by a comparison of grey-scale values of contiguous sequences of pixels (measurement zones 45) of the preceding image and .of the current image, and in that the following steps are performed: a) definition of size, position and type of the measurement zones in the image area, b) setting of the optimum exposure, c) extraction and storage of the reference values of each measurement zone and of the current exposure parameters of the camera d) continuous operation with regulation in adjustable time intervals.
2. Method for regulating the exposure according to Claim 1, characterized in that grey-scale value histograms are created from the grey-scale values of the measurement zones and assessed by means of a crosscorrelation and the correction factor is calculated from this .
3. Method for regulating the exposure according to Claim 1 or 2, characterized in that the correction factor is calculated as an average value over all the measurement zones.
4. Method for regulating the exposure according to one of the preceding claims, characterized in that each individual correction factor is checked with regard to its usability for the method by criteria for its plausibility being prescribed.
5. Method for regulating the exposure according to Claims 1 to 4, characterized in that measurement zones are evaluated for monitoring overdriving, grey-scale value histograms of the measurement zones being checked with regard to criteria for the presence of overdriving.
6. Method for regulating the exposure according to Claims 1 to 5, characterized in that the regulation has 5 the following steps: a) selection of the last recorded image or recording of a new image and extraction of the measurement data, b) comparison of the data of the current image with the stored reference data, determination of the correction value, c) . checking of the criteria with respect to overdriving, d) changing of the camera parameters if the magnitude of the correction value exceeds an adjustable value.
