DE19952832A1 - Video-assisted accident/traffic documentation system, stores defined amount of image data representing period including trigger point statically when triggering device outputs control signal - Google Patents

Abstract

The system has one or more electronic cameras (1.0) recording continuously when operating, at least one connected electronic control device (1.1) with at least one memory (1.3) for storing image data over a defined period at a suitable time interval and suitably organized an for subsequent organized overwriting. The time for which data remain in memory varies with proximity to the current time. When a triggering device outputs a control signal (1.5) a defined amount of image data representing a time section including the trigger point are stored statically. An Independent claim is also included for a method of documenting an accident and traffic situation.

Description

The invention relates to a video-based accident and traffic documentation system.

Due to the enormous increase in traffic density in the past few years, for every road user a high likelihood of being at fault Traffic accident involving more or less serious outcome. In many cases where no witnesses are available and the opposing side is not is willing to admit their fault, it is for the injured or forced Road users are difficult to avoid a legal dispute or in Litigation to make the culpable arrest of others believable. This condition often causes complex legal proceedings. If the opponent does not succeed To prove guilt, the injured party must in addition to his direct damage also assume the procedural costs.

As a result of the lack of opportunity in the past, accident events are conclusive to document, there is an increasing tendency among accident-causing parties basically the admission of culpable behavior and thus the responsibility for to pass on one's own misconduct in road traffic at the expense of the injured party. At On the one hand, this leads to the feeling of defenselessness and Legal uncertainty in road traffic, on the other hand to the desire to be involved in an emergency appropriate evidence to enforce your own claims.

So-called black box systems are one way of obtaining such evidence. These systems record relevant data from which the course of the accident is reconstructed can be.

Such an accident data recorder and a method for analyzing the accident occurrence is known from the patent DE 42 21 280 C2. The system described records video images recorded with a video camera in an electronic memory which is organized as a ring memory. In the event of a triggering event, the storage of the video data in the ring buffer is interrupted after a defined period of time in order to be read by a reading device and to be used for evaluating the accident. In addition to the video data, other data relevant to the reconstruction of an accident, such as driving parameters (date, time), vehicle operating status data and driver status data, are stored. The image data can be stored in a location-equidistant manner, that is, as the vehicle speed increases, the time interval between the recorded images decreases. Another possibility described is the time-equidistant recording, ie the time interval between the recorded images remains the same. The required storage space results from the formula

Memory requirement = image size. Recording time. Recording frequency

In order to get more detailed information on the course of the accident in the event of an accident, the patent describes a method which, after the triggering event Recording frequency increased.

The invention has for its object an electronic accident and Traffic documentation system to improve in that the same Storage requirements, vehicle type-independent usability and minimized installation effort  the meaningfulness of the recorded image sequence for the reconstruction of an accident is clear can be increased.

This object is achieved according to the invention for an accident and Traffic documentation system with the features specified in claim 1.

The accident and traffic documentation system consists of one or more electronic cameras permanently recording, at least one connected control electronics with at least one memory designed in this way is that in this over a predetermined period of time the image data in a suitable time interval and can be saved in a given organization and can then be overwritten in a given organization, the Dwell time of the recorded image data in the memory in relation to the temporal proximity is of different duration at the current time, and at least one Tripping device which is designed such that it is dependent on a triggering Moments sends a control signal to the control electronics, the control electronics and the digital memory are further designed such that depending on the control signal of the triggering device after a predetermined period of time Image data, which represent a time segment comprising the triggering moment, in the Memory are stored statically.

The different recording densities ensure that the available memory is used optimally: a very high temporal resolution in close proximity to the triggering event and a lower resolution in a larger temporal distance. The linear relationship between temporal resolution and recording duration in the temporal region before the triggering event, which is known in previously known systems, can be broken, as the distribution of the recording density shown in FIG. 2 shows. Here, with approximately the same content of relevant information, a reduction in the required memory by a factor of 3.6 can be achieved, which is a significant advantage especially when considering the high cost share of the memory in the total costs of an accident documentation system. Either an event called by the driver (push of a button, voice input) or an automatically generated event (e.g. by means of an acceleration sensor), which indicates an accident or an abnormal driving situation, can serve as the triggering event.

Another advantageous embodiment of the invention is that the Memory management is designed such that at least two different ones Record densities are defined and the largest record density in the temporal Closeness at the current time is greater than further away.

The different recording densities serve to reduce the memory requirement with approximately the same content of relevant information. In the simple case there are two Areas with different recording density are conceivable, but it can also be one completely discrete distribution of the recording density depending on the temporal Distance at the current time.

Another advantageous embodiment of the invention is that the Memory management is designed such that the different recording densities are achieved through organized storage.

The storage can take place in a conventional, linearly addressable memory. In contrast to a ring memory, the image data are not stored linearly one after the other, but are distributed in the memory, the data being managed using a table of pointers to memory areas. A newly arriving picture is written to the place of the now "least important" picture when the memory is already filled. A method for determining the "least important" image is given in the description of FIG. 3.

Another advantageous embodiment of the invention is that the stored image data are stored compressed. A changing over time Recording density can also be done with compressed image data. With lossy (MPEG, lossy JPEG) compression of the image data has to be considered that relevant information can be lost (e.g. an indistinctly recognizable License plate of a fugitive) and thus the benefits and the The usability of the recorded data may be significantly restricted. Therefore if possible, lossless compression is preferable.

Furthermore, protection for a system according to one of claims 5, 11, 12, 13 and 14 coveted. The features of claim 5 are also independent of claims 1 to 4 worthy of protection (possibly divisional registration).

Another advantageous embodiment of the invention is that at least two Cameras offset from one another with overlapping image areas are provided and that one Evaluation device is provided, which from the recorded image data of the overlapping image area the movements of vehicles and road users involved deduces. The existence of spatial information results in a largely autonomous accident and traffic documentation system, the recording of Operating data of the vehicle such as speed, steering angle etc. is not essential required. This also eliminates the need for own sensors or the dependency from connection to in-vehicle bus systems, whose protocols mostly depend on the manufacturer and are not disclosed. Thus, the accident and Traffic documentation system built into every vehicle without major adjustments only a signaling of the triggering moment is required.

Furthermore, the invention relates to a method for documenting an accident and traffic situation. The process is characterized by the following steps:

• a) continuous recording of image data from one or more electronic cameras,
• b) transmission of the image data by means of control electronics in a limited digital Storage,
• c) storing the image data at a predetermined time interval and in one given organization,
• d) overwriting the image data from a predetermined amount of data in the memory in a predetermined temporal organization in the memory, the storing of the image data in Dependence on the temporal proximity to the current point in time is different densities.
• e) Static storage of at least part of the image data due to a triggering Moments, where the triggering moment is represented by that represented by the image data Period is included.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that the image data in at least two groups different recording density with a defined group size.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that an image of a group with lower  Recording density belongs to a group with a higher recording density can belong.

Another advantageous embodiment of the method for documenting an accident and traffic situation consists in the fact that primarily the pictures with greater temporal Distance to the current point in time, depending on the Organization of the memory individual older image data over a predetermined period be held in memory longer than others.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that the recorded images with a film temporally equidistant images and one of the highest recording density Refresh rate can be summarized, with a reduced Recording density created gaps by adding duplicates of a temporally neighboring image. The time-synchronous representation thus obtained can used to visualize an accident or a traffic situation become.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that the offset of at least two to each other arranged cameras and representing an overlapping image area recorded images assembled into a stereoscopic film with equidistant images become. In this way, a spatial impression can also be conveyed during the visualization which is the reconstruction of an accident or a traffic situation facilitated.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that the images recorded by a camera with Methods of digital image processing are edited with the aim for Get relevant information about the course of the accident.

Another advantageous embodiment of the method for documenting an accident and traffic situation is that of one or more at least two offset from one another and representing an overlapping image area Cameras recorded images with methods of digital image processing with the aim processed to receive information relevant to the course of the accident. By Stereoscopic procedures can give precise information about speed and Position relative to fixed and moving objects in the recorded video images be made. The big advantage is that the camera has its own sensors for Physical quantities (e.g. speed, rotation) are no longer required. Only the determination of a triggering moment must be made using suitable sensors (e.g. determination of a limit value for acceleration being exceeded). The Connection to an in-vehicle bus system (e.g. CAN) with the aim of logging the physical parameters transmitted on the bus can be omitted. This is a special advantage because the protocols used on the in-vehicle buses manufacturer-specific and generally not disclosed. A use of the in-vehicle buses to preserve information about physical vehicle variables therefore always requires a manufacturer-specific software variant.

Fig. 1 shows the block diagram of the accident and traffic documentation system.

Fig. 2 shows an example of the recording density of image data depending on the current time.

Figure 3 shows the flow diagram of a possible implementation of memory management.

Fig. 4 shows the change over time of a 12 image detecting memory with two different record densities over a period of 45 incoming images, it being understood by an initial empty memory.

Fig. 5 shows an example of recorded data, which are available after a triggering event in the memory.

In Fig. 1 the block diagram of the accident and traffic documentation system is shown. A continuous stream of image data is generated by means of a camera 1.0 , the path 1.2 of which is determined by means of control electronics. The control electronics ensure that the image data are written to a suitable place in part of the image memory 1.3 . If a triggering event 1.5 occurs, a limited amount of incoming image data is stored in a previously reserved part of the image memory. If this is filled, the content of the entire image memory is retained, and newly arriving image data are discarded. The recorded image data are available for reading and can be used to reconstruct the course of the accident.

In FIG. 2, an example of two different recording densities of image data depending on the current date is given. Group 1 2.1 records over a period of 20 seconds with a recording density of 30 frames / second. Group 2 2.2 records over a period of 100 seconds with a recording density of 4 frames / second. This results in a memory requirement of (20.30 + 100.4) = 1000 images. Despite the memory of only 1000 images, statements about the course of the accident can be made with a very high temporal resolution in the vicinity of the triggering event, whereby a relatively long time range before the accident is also documented. If an accident documentation system with the previously known ring buffer were to display the same time period with the same high temporal resolution, an image memory would be required, which is 120. 30 = 3600 images can be saved. Since the image memory is one of the most expensive components of a video-based accident documentation system, there are significant cost advantages with the solution according to the invention.

FIG. 3 shows a possible implementation of the memory management algorithm using a flow chart. It is assumed that three groups with different recording densities are selected, the highest storage density corresponding to the rate of the incoming images. After the start 3.1 there is a wait for a new arriving picture 3.2 . The next step is to check whether a triggering event 3.3 has occurred in the meantime. If this is not the case, a check is carried out to determine whether all the storage locations in the intended storage area have already been used 3.4 . If there is still space in the memory, the current image is saved in a free memory location 3.5 , the program jumps back and waits for a newly arriving image 3.2 . If the check 3.4 shows that the memory is already occupied, a memory location must be selected which can be overwritten. For this purpose, it is checked whether the addition of the image fills the number of members of the group of images with the lowest storage density beyond the maximum predetermined number 3.6 . If this is the case, the oldest picture of the group with the lowest storage density is released for overwriting with the current picture 3.7 , after writing the current picture the program jumps back and waits for a newly arriving picture 3.2 . If the maximum number of members of the group of images with the lowest storage density is not exceeded by adding the newly arriving image 3.6 , it is checked whether the addition of the image exceeds the number of members of the group of images with the second lowest storage density above the maximum specified number replenishes 3.8 . If this is the case, the oldest picture of the group with the second lowest storage density is released for overwriting with the current picture 3.9 , after writing the current picture the program jumps back and waits for a newly arriving picture 3.2 . If the maximum number of members of the group of images with the second lowest storage density is not exceeded 3.8 by adding the newly arriving image, then the oldest image of the group with the highest storage density is overwritten, the program jumps back and waits for a newly arriving image 3.2 . If the triggering event occurs 3.3 , then no more images are stored in the previous part of the image memory, but the newly arriving images are stored in this until a previously reserved part of the memory is filled 3.11 . In the beginning, the filling can be done with a high recording density; from a certain time distance from the triggering event, the recording density can be reduced, ie individual images are omitted during storage. The content of the image memory with the image data recorded before and after the triggering event remains statically and is available for reading out 3.12 .

In FIG. 4, the temporal change of a 12 image detecting memory block is shown at two different recording densities over a period of 45 incoming images 4.2 4.1, it being understood by an initial empty memory. The following recording densities were determined:
Group 1: 5 pictures with recording every 5th picture
Group 2: the remaining storage space with recording of each picture.

The first 12 incoming images are sequentially stored in the initially empty memory written. Pictures 5 and 10 are members of group 1 (modulo 5) recognized. The incoming 13th picture finds a full memory. Therefore, after the procedure described above checked whether adding the 13th picture the number the members of group 1 would raise above the maximum value (5). This is not the one Case, therefore the oldest member of group 2 (picture 1) is overwritten. That new incoming picture 14 overwrites picture 2 etc. With the arrival of picture 25 the maximum number of members of group 1 reached. The incoming picture 30 overwrites the oldest picture in group 1, picture 5, as a member of this group overwrites picture 35 then picture 10, picture 40 picture 15 etc. all not belonging to group 1 Pictures overwrite the oldest picture in group 2. This results in filling of the memory and filling all groups a stable allocation of the memory, there are 5 pictures from group 1 exist, the remaining pictures belong to group 2, members of group 1 being members of group 2 for a certain period of time can (e.g. Fig. 40).

In Fig. 5 there is shown an example of how image data associated with different recording density for accident reconstruction available. The two groups of FIG. 2 were adopted with a recording rate of 30 frames 5.1 and 4 frames 5.2 . After a triggering moment, these two groups are frozen and the newly arriving image data is saved in two further groups: Group 3 5.3 with a storage density of 15 frames / second and Group 4 5.4 with a storage density of 2 images / second. With the data recorded here, it is possible to reconstruct the course of the accident over a period of 260 seconds with a very high temporal resolution in close proximity to the triggering event. The numerical values given are only examples, they can be varied according to the requirements.

Claims (14)

1. Accident and traffic documentation system with one or more electronic im Use of permanently recording cameras, at least one connected Control electronics with at least one memory which is designed such that in the latter over a predetermined period of time the image data in a suitable temporal Distance and can be saved in a given organization and then in a given organization can be overwritten, the length of stay of recorded image data in the memory in relation to the temporal proximity to the current Time is of different duration, and at least one triggering device is designed such that it is a function of a triggering moment Sends control signal to the control electronics, the control electronics and the digital Memory are further designed such that the Tripping device after a predetermined period of time to a predetermined number Image data covering a period of time that includes the triggering moment represent statically stored in memory. 2. System according to claim 1, characterized in that the memory management is designed such that at least two different recording densities are defined and the greatest recording density in the temporal proximity to the current one Time is greater than further away. 3. System according to at least one of claims 1 to 2, characterized in that the memory management is designed such that the different Record densities are achieved through organized storage. 4. System according to at least one of claims 1 to 3, characterized in that the stored image data are stored in compressed form. 5. System according to at least one of claims 1 to 4, characterized in that at least two cameras offset from one another with an overlapping image area are provided and that an evaluation device is provided, which from the recorded image data of the overlapping image area the movements involved vehicles and road users. 6. Procedure for documenting an accident and traffic situation, characterized by the following steps:

• a) continuous recording of image data from one or more electronic cameras,
• b) transmission of the image data by means of control electronics into a limited digital memory,
• c) storing the image data at a predetermined time interval and in a predetermined organization,
• d) Overwriting the image data from a predetermined amount of data in the memory in a predetermined temporal organization in the memory, the storage of the image data depending on the temporal proximity to the current point of time being carried out with different densities.
• e) Static storage of at least part of the image data on the basis of a triggering moment, the triggering moment being comprised by the time period represented by the image data.

7. The method according to at least one of claims 1 to 6, characterized in that that the image data in at least two groups of different recording density with a defined group size. 8. The method according to at least one of claims 1 to 7, characterized in that that an image belonging to a group with a low recording density also can belong to a group with higher recording density.   9. The method according to at least one of claims 1 to 8, characterized in that primarily the images with a greater temporal distance from the current point in time are overwritten, depending on the organization of the memory individual older image data is held in memory for a predetermined period of time be than others. 10. The method according to at least one of claims 1 to 9, characterized in that a new arriving picture with already filled memory the picture with the least importance overwrites in memory, the importance of an image from time interval of the recording at the current time and from the affiliation to one or more groups with different recording density. 11. The method according to at least one of claims 1 to 10, characterized in that that the recorded images into a film with temporally equidistant images and a refresh rate corresponding to the highest recording density be summarized, with a reduced recording density created gaps by adding duplicates of a temporally adjacent Image to be filled. 12. The method according to at least one of claims 1 to 11, characterized in that that at least two staggered and one cameras representing overlapping image area stereoscopic film can be mounted with equidistant images. 13. The method according to at least one of claims 1 to 12, characterized in that that the recorded images with methods of digital image processing with the aim processed to receive information relevant to the course of the accident. 14. The method according to at least one of claims 1 to 13, characterized in that that at least two staggered and one overlapping image-representing cameras with recorded images Methods of digital image processing are edited with the aim for Get relevant information about the course of the accident.