TWI408612B - Method and system for dynamically and simultaneously determining the relative relation between moving objects - Google Patents

Abstract

The present invention provides a method for dynamically and simultaneously determining the characteristic relation between moving objects, which comprises steps of: establishing a relation between the image acquiring distance and a image characteristic associated with an object; acquiring an image with respect to a moving apparatus having the object disposed thereon; finding a image area corresponding to the object and determining an image characteristic associated with the object within the image area. Finally, a step is performed to determine the image acquired distance with respect to the object according to the established relation. By means of the foregoing determining method, the present invention further provides a determining system, disposed on the moving apparatus, for acquiring the image of the other moving apparatus, thereby determining the relative relation between the two moving apparatus for being the basis of the safety measure.

Description

Dynamic instant relative relationship identification method and system

The invention relates to an image recognition technology, in particular to a dynamic instant relative relationship identification method and system for judging the relative relationship between two objects according to image features.

Although there are many kinds of driving recording devices on the market, they are limited by the size of the memory, and can only store the latest minutes to one or two hours of data. However, in the event of a car accident or accident, driving may not be able to turn off the function of image recording (such as coma, medical treatment, or inability to move), causing the recording device to continue to operate, and to cover the data when the accident occurred. The image at the most critical moment disappeared and could not be used as a reference. If the driving recorder is turned on with the movement of the vehicle, the recorder cannot provide key images when stopping waiting for a red light or parking on the road; especially when the roadside parking is accidentally hit, most of the perpetrators Will escape, at this time the key image will become an important basis for the claim and the circumstances of the anecdote.

In the conventional technology, for example, Japanese laid-open application JP2009032143 discloses a driving recorder for reducing the risk of disappearing of evidence data in the event of an accident, in which the image signal in front of the vehicle is captured during the travel of the vehicle, and the image signal is converted. The animation image information is recorded in the memory. The technique divides an animated image into a plurality of still images, and then, when an accident occurs, a still image corresponding to the occurrence time point can be sent out using a wireless network.

In addition, the Republic of China Patent Publication No. TW200409051 also discloses a technique for recording images to ensure the authenticity of data in the event of a car accident. In this technique, the camera is separated from the distance detection, and the detector is used to detect objects around the vehicle. If the object is too close, the camera is started to start recording.

The invention provides a dynamic instant relative relationship identification method and system, which utilizes the relative relationship between image detection and a movable device in a circumferential environment, and activates a corresponding mechanism according to the relative relationship to alert the driver or to a situation that may cause an accident. Store images.

The invention provides a dynamic instant relative relationship identification method and system, which uses an image capturing device to capture images of vehicles in a surrounding environment, and detects the characteristics of the license plate in the image, and estimates the speed of the vehicle according to the license plate characteristics. At the same time, the environmental detection camera also detects environmental information such as traffic lights and speed limit signals to determine whether it is likely to cause harm to my car.

The invention provides a dynamic instant relative relationship identification method and system, which is recorded by visual images, does not need to rely on other additional equipment, detects the state of distance, speed or acceleration with other vehicles, and instantly records images, the present invention The threshold value of the recorded parameter can be adjusted according to environmental dynamics (for example: traffic light number, distance between two workshops, speed and acceleration, or speed limit).

In an embodiment, the present invention provides a dynamic instant relative relationship identification method, which includes the following steps: establishing an image capture distance of an image capture device and a feature relationship of an image feature of a feature; Taking an image of a movable device having the feature; detecting an image region of the feature in the image; determining an image feature for detecting one of the features in the image region; and detecting the image according to the detected image The feature determines the relative relationship between the image capturing device and the feature in the feature relationship. In another embodiment, a mechanism for alerting or recording can be initiated based on the relative relationship.

In another embodiment, the present invention further provides a dynamic vehicle instantaneous relative relationship identification method, which comprises the steps of: establishing a feature relationship between an image capturing distance of an image capturing device and an image feature of the license plate; Taking an image of a first vehicle having a license plate; detecting an image area of the license plate in the image; determining an image feature of one of the license plates in the image area; and detecting the image according to the detected image The feature determines the relative relationship between the image capturing device and the license plate in the feature relationship. In another embodiment, a mechanism for alerting or recording can be initiated based on the relative relationship.

In another embodiment, the present invention further provides a dynamic instant relative relationship identification system, which includes: a database that stores a feature relationship between an image capture distance and an image feature of a feature; An image capturing device that captures an image of a movable device having the feature in an external environment; a control unit coupled to the database and the image capturing device, the control unit Detecting an image area of the feature in the image, and then determining, in the image area, detecting image features of one of the features, and determining the image capturing device and the feature according to the detected image feature according to the detected image feature The relative relationship between the two.

In order to enable the reviewing committee to have a further understanding and understanding of the features, objects and functions of the present invention, the related detailed structure of the device of the present invention and the concept of the design are explained below so that the reviewing committee can understand the present invention. Features, detailed descriptions are as follows:

Please refer to FIG. 1 , which is a schematic flowchart of a first embodiment of a dynamic instant relative relationship identification method according to the present invention. In this embodiment, the method includes the following steps: First, in step 20, a feature relationship between an image capturing distance of an image capturing device and an image feature of a feature is established. In this step, a characteristic relationship is first established by utilizing the characteristics that the image features of the feature image change with the distance from the image capturing device. Taking the feature as the license plate as an example (but not limited thereto), as shown in FIG. 2A, the figure is a schematic diagram of the relationship between the image capturing device and the license plate position. The license plate 90 is disposed on the vehicle 91. An image capturing device 92 is disposed in front of the vehicle 91. The height of the license plate is set to H (90 cm), the distance from the license plate is D (80 cm to 300 cm), and the photographing angle is θ ( 17 degrees). As the image capture distance changes, the size of the vehicle 91 for the same size captured image 98 will also change, as shown in FIG. 2C and FIG. 2D. The image capturing distance of Figure 2C is greater than the image capturing distance of Figure 2D. The relative relationship between the image capturing distance and the feature is shown in Table 1 below. As shown in FIG. 2B, the license plate width (top) in Table 1 is W ₁ , the license plate width (bottom) is W ₂ , the license plate height is h, and the unit is pixel (pixel). The image features in step 20 are the license plate width, license plate height, area, average width or average area in Table 1.

According to the characteristic relationship of Table 1, the results shown in FIG. 3A and FIG. 3B can be obtained, wherein FIG. 3A represents the relationship between the image capturing distance and the average width, and FIG. 3B represents the image capturing distance and the average area. Relationship lines. According to the curve of FIG. 3A or FIG. 3B, it can be used to compare the image feature size of the license plate under different distances of different images. As for the average width and the average area, since the image capturing device has an inclination angle θ with respect to the license plate, the upper width and the lower width of the license plate may differ. Whether or not to use the average width and the average area as image features depends on the needs of the user. In addition, after having the curve of FIG. 3A or FIG. 3B, an algorithm for mathematical fitting can be used to establish an equation about the curve. It is a customary technique to establish a fitting relationship according to the data points, and no further description is made here.

Returning to FIG. 1, after step 20, an image of one of the movable devices is captured by step 21, and the movable device has the feature. After the relative relationship established in step 20, an image of any animal having the feature can be captured, and an image between the image capturing device and the feature is determined according to the image feature of the feature in the image. distance. Therefore, after the image of the movable device is obtained in step 21, the image area of the feature is detected in the image in step 22. In this step, the image captured in step 21 is mainly identified to find an image area corresponding to the feature. In step 22, the candidate block is mainly searched according to the boundary strength, and the boundary of the candidate block is taken as the boundary corresponding to the feature image.

After determining the image area, step 23 is then used to determine image characteristics for one of the features in the image area. In this step, after the position of the image area is obtained, the width, average width, height, area, or average area of the feature area can be obtained by using image processing as the detected image feature of the feature image. . After the detected image feature is obtained, step 24 is performed, and according to the detected image feature, the relative relationship between the image capturing device and the feature is determined in the feature relationship. In step 24, the equation corresponding to the relative relationship curve established in step 20 is used, and the detected image feature determined in step 23 is brought into the equation, and the corresponding image capturing distance can be obtained. In addition, the relative relationship in step 24 may be a relative speed or a relative acceleration between the feature and the image capturing device, in addition to the image capturing distance. As for judging the relative velocity and acceleration of the feature, it is only necessary to obtain a change in the distance according to the feature relationship at two time points, and the relative velocity can be calculated. Similarly, the relative speed difference between the two time points is the relative acceleration.

Referring to FIG. 4, the figure is a schematic flowchart of a second embodiment of a dynamic instant relative relationship identification method according to the present invention. The method 3 includes the following steps. First, in step 30, a feature relationship between an image capturing distance of an image capturing device and an image feature of the license plate is established. The method of this step is the same as that of step 20, except that the feature is a license plate. Because the size of the license plate is fixed, the relationship between the image capturing distance and the characteristic size, shape or area of the license plate can be established in advance, as shown in step 20. As mentioned, I will not repeat them here. Then, in step 31, an image of one of the first vehicles is captured, and the first vehicle has a license plate. As shown in FIG. 5A, the image capturing device 92 is disposed on a second vehicle 910. In the embodiment, the image capturing device 92 is disposed behind the second vehicle 910 to capture the rear first. The image of the vehicle 911 is an image provided in the rear of the second vehicle 910 to capture the rear vehicle, but may be disposed in front of the vehicle (as shown by the first vehicle 911) or on the left and right. The position of the side, which may be determined according to actual needs and not limited to the embodiment of FIG. 5A.

Returning to FIG. 4, after step 31, the image area of the license plate is detected in the image by step 32. Please refer to FIG. 5B, which is a schematic diagram of the vehicle image taken at a certain distance. In the image 97 of FIG. 5B, there is an image of the license plate, so the candidate block 93 is searched for the image area according to the boundary strength by the image processing algorithm. Then, in step 33, it is determined in the image area that the image feature is detected with respect to one of the license plates. In this step, after obtaining the image area of the license plate, the size feature of the license plate is detected in the image area, and the detectable features include the width, height or area of the license plate as the detection. The image feature, if the image capture device has an angle of inclination with respect to the captured license plate, as shown in FIG. 2A, the average width or the average area may be used as the detected image feature of the license plate.

Then, in step 34, the relative relationship between the image capturing device and the license plate is determined according to the detected image feature in the feature relationship. Since the image capturing distance of the image capturing device and the characteristic relationship between the license plates have been established in step 30, according to the relationship, the equation corresponding to the characteristic relationship curve can be fitted by the mathematical equation. According to the fitted equation, the image capturing distance corresponding to the image feature can be obtained as a result of the relative relationship of step 34. Since the image capturing device is disposed on the second vehicle, the distance obtained in step 34 can be used as the distance L between the two vehicles in FIG. The relative relationship in step 34 may be a relative speed between the two vehicles or a relative acceleration, in addition to the image capturing distance. As for the relative speed or the relative acceleration information, the relative speed between the two vehicles can be known from the distance difference ΔL obtained by the time difference ΔT between two different time points, and the relative acceleration can be obtained from the speed difference between the two different time points. Since the vehicle speed of the second vehicle is known, the vehicle speed of the rear vehicle (first vehicle) can be obtained according to the relative speed.

With the relative relationship between the vehicles, step 35 can be performed to determine whether it is a critical moment based on the relative relationship, and if it is the critical moment, a processing operation is performed in step 36. Otherwise, go back to step 31. The key moment in step 35 is whether the relative relationship exceeds a specific value or a specific range (distance, relative speed or relative acceleration), for example, if the distance between the first vehicle and the second vehicle is less than the distance of the safety brake The point in time is the meaning of the key moment. As for the specific value or the specific range, it can be set in advance and stored, or it can be automatically adjusted according to the environmental state. In step 36, as shown in FIG. 5A, the processing action may be an alert message to notify the driver of the second vehicle 910 or control the image capturing device 92 in the second vehicle 910 to capture the image of the first vehicle 911. (Static or dynamic recording) The image of the first vehicle 911 is then stored as an important proof of future accidents. In addition, warnings and recorded images can also be performed simultaneously.

In addition, after step 32, step 37 may be synchronously performed, and an image recognition program is used to identify the license plate content in the image area, and the identification content is stored as a basis for assisting the judgment. Please refer to FIG. 6 , which is a schematic diagram of the process of recognizing license plate image content according to the present invention. The identification method 4 first provides, in step 40, a database in which a plurality of known standard sample images are created. As shown in Fig. 7A, the figure is a schematic view of the size of a known sample image. The size of the known sample image 5 is determined by the user, for example, 130 (pixels) x 130 (pixels), but not limited thereto. A standard image area 50 is formed on the pixels within the known sample image 5. The standard image area 50 is composed of a plurality of pixels 500 and 501 to form characters, numbers, characters or patterns to be represented by the known sample image. Referring to FIG. 7B, in the present embodiment, the number 1 is used to describe a suitable gray level value for each pixel 500 and 501 in the region of the known sample image 5 to form a standard image area 50. And outline the shape of the number 1. A particular pixel 501 (pixel of the shaded area) is then determined within the standard image area 50 to give a particular weight. The value of the gradation value and the weight value may be determined according to requirements, that is, each weight value may be different or the same. In this embodiment, the weight value is a positive value. The gray value and the weight value of each of the pixels 500 and 501 in the standard image area 50 are the first feature value.

As shown in Figure VIIC, a non-standard image area 51 is determined within the known sample image. The non-standard image area 51 is a character content indicating that the character formed by the standard video area 50 is easily misidentified. For example, the number "1" is easily mistaken for the English letter "I" or "L" or even the letter "E" in the image. Therefore, the relevant pixel position 510 (pixel of the dot area) that may be misidentified and the appropriate gray value and weight are given as the second feature value of the corresponding pixel 510. In the present embodiment, the position of the pixel 510 constituting the non-standard image area 51 can be determined based on characters, numerals, characters, and the like which are easily misidentified by the standard image area 50, and there is no rule. The value of the gray value and the weight may be determined according to requirements. In this embodiment, the weight of 51 in the non-standard image area is a negative value.

As shown in Fig. 7D, the figure is another schematic diagram of a known standard image. This figure is a known sample image 5a established according to the number 0. The known sample image 5a also has a standard image area and a non-standard image area. The pattern formed by each pixel in the standard image area is the number "0". Similarly, the pattern formed by each pixel in the non-standard image area represents a character that is easily misidentified by the number “0”, for example, the letter “Q” or the number “8”. The way to establish a non-standard area of a known standard image can be handled by an image software, such as a small painter, but not limited thereto. The foregoing is a process for generating a so-called standard image according to the present invention, and sequentially creates known sample images represented by different characters or numbers according to the foregoing manner, for example, 0~9, A~Z, and a~z, and stored in the database. .

Returning to FIG. 6, proceeding to step 41, a feature image is captured in the reconstructed image. For example, taking FIG. 8A as an example, the image area 94 corresponding to the license plate is taken out from the image of the vehicle, and the area corresponding to each unidentified character is the feature image. The feature image is then normalized to adjust the size and angle of the feature image such that the size of the feature image is consistent with the size of the known sample image for subsequent recognition. In step 41, the captured feature image 95 is the first code text of the vehicle identification information. Then performing step 42 to perform a calculation on the third feature value of each pixel in each of the feature images and the first feature value or the second feature value corresponding to each pixel of the plurality of known sample images in the database. To obtain the feature image, one of the plurality of known sample images has a similarity value.

Please refer to FIG. 8B, which is a schematic diagram of the feature image 95. Using the feature image, each known sample image can be calculated to obtain a corresponding similarity value C _uv . The calculation method is a formal correlation comparison method or an approximation comparison method, but is not limited thereto. Taking the formal correlation comparison method as an example, the calculation formula is as shown in the following formula (1). The normalized correlation matching is mainly to calculate the relationship between the feature image and the known sample image. The standard deviation of the gray value in each image is regarded as a vector and multiplied by the weight. In order to determine which is the best matching position, the normalized cross-correlation coefficient is between -1 and 1, the closer to 1 is, the higher the similarity is; when _Cuv is the highest, it is the best matching position.

Where u _i is the gray value of each pixel in the known standard image, and v _i is the gray value of each pixel in the feature image. Is the grayscale average of all pixels in the known standard image. Is the grayscale average of all pixels in the feature image. The w _i is the weight represented by the pixels in the standard image area and the non-standard image area in the known sample image, and the pixels of other areas have a weight of 1.

Each pixel of FIG. 8B is calculated with each pixel of the known sample image according to equation (1). For example, the image of Figure 8B and the known sample image of Figure 7C (representative number 1) and the known sample image of Figure 7D (representative number 0) are calculated separately, and the characteristic image of Figure 8B can be obtained. Figure 7C and Figure 7D have similarity values C _uv . Returning to FIG. 6, after obtaining the similarity value, all the characters in the reconstructed image 95 are successively extracted into feature images by steps 43 and 44, and then step 42 is repeated for comparison. Then, in step 45, a plurality of similarity values generated by comparing the feature image with the plurality of known sample images are summarized. In this step, the similarity values can be sorted, and the most likely identification results are sorted to the lowest result. Finally, in step 46, the plurality of similarity values are sorted to output a plurality of possible identification comparison results, and the license plate having seven codes is taken as an example. As shown in FIG. 8A, the figure is a schematic diagram of the license plate image. After the process of the method 4 is identified, the sort result of FIG. 9 can be obtained. In Figure IX, a total of four possible outcomes are output, each of which represents a possible combination of characters for the license plate content. Each code of the first possible result has the highest degree of similarity and is then sequentially arranged to form the second, third, and fourth possible outcomes. Taking the first possible result as an example, the possible license plate analyzed is 0695-0A, wherein the similarity after the calculation of the first code “0” is 52, and the similarity after the second code “6” is calculated. 72, the third code "9" has a similarity after the calculation of 67, the fourth code "5" after the calculation has a similarity of 72, the fifth code is "-", the sixth code "0" is passed The similarity after the calculation is 63, and the similarity after the calculation of the 7th code "A" is 76. Of course, the user can also determine other possible license plate number combinations for the relevant units to confirm according to the results of FIG.

Please refer to FIG. 10, which is a schematic diagram of an embodiment of a dynamic instant relative relationship identification system of the present invention. The identification system 6 includes a database 60, a first image capturing device 61, a control unit 62, and a warning unit 63. The identification system 6 is disposed on a mobile vehicle, and the mobile vehicle can be a wheel type vehicle, for example, a two-wheel or four-wheel type vehicle (such as the second vehicle 910 in FIG. 5A). But not limited to this. The database 60 stores a feature relationship between the image capturing distance and an image feature of a feature, that is, the information or relationship curve established by the steps 20 and 30 in the storage chart 1 or FIG. function. The database 60 is constructed as a storage medium, such as a magnetic storage medium, an optical storage medium, or a memory. The first image capturing device 61 is configured to capture an image of a movable device having the feature in an external environment. In the embodiment, the movable device is the first vehicle in FIG. 5A. 911, and the feature is the license plate 90, but not limited thereto. It can be applied as long as it is a movable object and has a feature that needs to be recognized and recorded. The control unit 62 is connected to the database 60 and the first image capturing device 61. The control unit 62 performs the steps of steps 22-24 of FIG. 1 or steps 32 to 34 of FIG. Detecting an image area of the feature in the image, and then determining, in the image area, detecting image features of one of the features, and determining the image capturing device and the feature according to the detected image feature according to the detected image feature The relative relationship between the two. The detailed steps are as described above and will not be described here.

After the control unit determines the relative relationship, the warning unit 63 may be activated to issue a warning to remind the driver or to capture the first image capturing device 61 according to the state of the relative relationship, that is, the determining manner in step 35. The image is recorded in the storage unit 64 in a static (single photo) manner or in a dynamic video recording manner. In addition, the control unit 62 can be connected to a second image capturing device 65, which can be disposed in front of the vehicle or on the left and right sides, so as to capture an external environment (for example, a speed limit) or a number. Image of Zhi (traffic light). The control unit 62 can further identify the image captured by the second image capturing device 65 to determine whether the warning unit 63 is used when the light is red or the speed of the moving vehicle approaches or exceeds the speed limit. Send a warning message. The control unit 62 determines, according to the traffic signal state and the speed limit number image, whether the light source is red or the speed of the movable device approaches or exceeds the speed limit, and the first image capturing device 61 The image captured by the image capturing device 62 is recorded in the storage unit 64. As a supporting evidence for the red light of the vehicle that is overspeeded or opposite or to be turned at the time of the accident.

However, the above is only an embodiment of the present invention, and the scope of the present invention is not limited thereto. It is to be understood that the scope of the present invention is not limited by the spirit and scope of the present invention, and should be considered as a further embodiment of the present invention.

2. . . Dynamic instant relative relationship identification method

20~24. . . step

3. . . Dynamic instant relative relationship identification method

30~37. . . step

4. . . Identification method

40~46. . . step

5. . . Sample image

50. . . Standard image area

500, 501. . . Pixel

51. . . Non-standard image area

510. . . Pixel

6. . . Identification system

60. . . database

61. . . First image capturing device

62. . . control unit

63. . . Warning unit

64. . . Storage unit

65. . . Second image capturing device

90. . . License plate

91. . . vehicle

910. . . Second vehicle

911. . . First vehicle

92. . . Image capture device

93. . . Candidate block

94. . . Image area

95. . . Feature image

97, 98. . . image

FIG. 1 is a schematic flow chart of a first embodiment of a dynamic instant relative relationship identification method according to the present invention.

Figure 2A is a schematic diagram showing the relationship between the image capturing device and the license plate position.

Figure 2B is a schematic diagram of the license plate image.

Figure 2C and Figure 2D are schematic diagrams of vehicle images taken from different image capturing distances.

Figure 3A represents the relationship between image capture distance and average width.

Figure 3B represents the relationship between image capture distance and average area.

FIG. 4 is a schematic flow chart of a second embodiment of a dynamic instant relative relationship identification method according to the present invention.

Figure 5A is a schematic diagram of the phase position relationship between two vehicles.

Figure 5B is a schematic diagram of the vehicle image taken at a specific distance.

FIG. 6 is a schematic diagram of the process of recognizing license plate image content according to the present invention.

Figures 7A through 7D are schematic diagrams showing the generation of known standard images.

Figure 8A is a schematic diagram of a reconstructed image and its feature image.

Figure 8B is a schematic diagram of a feature image.

Figure 9 is a schematic diagram showing the possible output results of the vehicle identification number according to the present invention.

Figure 10 is a schematic diagram of an embodiment of a dynamic instant relative relationship identification system of the present invention.

2. . . Dynamic instant relative relationship identification method

20~24. . . step

Claims (34)

A dynamic instant relative relationship identification method includes the steps of: establishing a feature relationship between an image capture distance of an image capture device and an image feature of a feature; capturing an image of a movable device, The movable device has the feature; the image area of the feature is detected in the image; the image feature is determined in one of the features in the image area; and the image is determined in the feature relationship according to the detected image feature The relative relationship between the device and the feature is captured. The dynamic instant relative relationship identification method as described in claim 1, wherein the relative relationship includes a distance and an image capturing distance to determine a rate of change of the distance with time. The method for identifying a dynamic real-time relative relationship as described in claim 1 further includes the step of initiating image storage when the relative relationship exceeds a specific value. The method for identifying a dynamic real-time relative relationship as described in claim 1 further includes a step of discriminating the content of the image area, wherein identifying the content of the image area further comprises the following steps: providing a a database having a plurality of known sample images, each of the known sample images having a standard image area and at least one non-standard image area, wherein the pixels in the standard image area respectively have a corresponding first characteristic value The pixels in the non-standard image area respectively correspond to a second feature value; at least one feature image is captured in the image region; and a third feature value of each of the pixels in each feature image is respectively associated with the complex number Performing a calculation on the first eigenvalue or the second eigenvalue corresponding to each pixel in the known sample image to obtain a similarity value of the plurality of known sample images corresponding to the feature image; And comparing a plurality of similarity values generated by the image to the plurality of known sample images; and ranking the plurality of similarity values The sequence output may be a plurality of identification alignment results. For example, the dynamic instant relative relationship identification method described in claim 4, wherein the calculus may be a formal correlation comparison method or an approximation comparison method. The dynamic instant relative relationship identification method according to claim 4, wherein each known sample image corresponds to a number or a character image. The method for identifying a dynamic real-time relative relationship as described in claim 4, wherein the first feature value and the second feature value are respectively a combination of a weight value and a gray value, and the third feature value is a gray level. value. The dynamic instant relative relationship identification method according to claim 1, wherein the image feature is a size, an area or a shape of the feature. The dynamic instant relative relationship identification method according to claim 1, wherein the detected image feature is a size, an area or a shape of the feature on the movable device. A method for identifying a real-time relative relationship of a dynamic vehicle, comprising the steps of: establishing a feature relationship between an image capturing distance of an image capturing device and an image feature of the license plate; and capturing an image of the first vehicle, The first vehicle has a license plate; the image area of the license plate is detected in the image; the image feature is determined in one of the license plates in the image area; and the image capture is determined according to the detected image feature in the feature relationship The relative relationship between the device and the license plate. The method for identifying a dynamic real-time relative relationship as described in claim 10, wherein the relative relationship includes a distance and an image capturing distance to determine a rate of change of the distance with time. The method for identifying a dynamic real-time relative relationship as described in claim 10, further comprising the step of initiating image storage when the relative relationship exceeds a specific value. The method for identifying a dynamic real-time relative relationship as described in claim 10, further comprising a step of discriminating the content of the image area, wherein the identifying the content of the image area further comprises the following steps: providing a a database having a plurality of known sample images, each of the known sample images having a standard image area and at least one non-standard image area, wherein the pixels in the standard image area respectively have a corresponding first characteristic value The pixels in the non-standard image area respectively correspond to a second feature value; at least one feature image is captured in the image region; and a third feature value of each of the pixels in each feature image is respectively associated with the complex number Performing a calculation on the first eigenvalue or the second eigenvalue corresponding to each pixel in the known sample image to obtain a similarity value of the plurality of known sample images corresponding to the feature image; And comparing a plurality of similarity values generated by the image to the plurality of known sample images; and ranking the plurality of similarity values The output may be plural kinds of identification than the results. For example, the dynamic real-time relative relationship identification method described in claim 13 may be a normal correlation comparison method or an approximation comparison method. The dynamic instant relative relationship identification method according to claim 13, wherein each known sample image corresponds to a number or a character image. The dynamic instant relative relationship identification method according to claim 13, wherein the first feature value and the second feature value are respectively a combination of a weight value and a gray value, and the third feature value is a gray level. value. The dynamic instant relative relationship identification method according to claim 10, wherein the image feature is the size, area or shape of the license plate. The dynamic instant relative relationship identification method according to claim 10, wherein the detected image feature is a size, an area or a shape of a license plate on the first vehicle. For example, the dynamic instant relative relationship identification method described in claim 10 further includes the steps of capturing the traffic signal state and the speed limit number image for identification. The method for identifying a dynamic real-time relative relationship as described in claim 19, further comprising the step of starting the image storage mechanism when the light is red or when the speed of the first vehicle approaches or exceeds the speed limit. The method for identifying a dynamic real-time relative relationship as described in claim 19, wherein the image capturing device is disposed on a second vehicle, and the dynamic instant relative relationship identification method further comprises: the light source is a red light or the A warning message is issued when the speed of the second vehicle approaches or exceeds the speed limit. A dynamic instant relative relationship identification system includes: a database storing a feature relationship between an image capture distance and an image feature of a feature; a first image capture device, the system Taking an image of the movable device having the feature in the external environment; a control unit connected to the database and the first image capturing device, wherein the control unit detects the image in the image The image area of the feature is then determined in the image area to detect image features of one of the features, and then the relative relationship between the first image capturing device and the feature is determined according to the detected image feature. For example, the dynamic instant relative relationship identification system described in claim 22, wherein the relative relationship includes a distance and an image capturing distance to determine a rate of change of the distance with time. The dynamic real-time relative relationship identification system described in claim 23, wherein the database further has a plurality of known sample images, each of the known sample images having a standard image area and at least one non-standard image area. The pixels in the standard image area respectively have a corresponding one of the first feature values, and the pixels in the non-standard image area respectively correspond to a second feature value. The dynamic real-time relative relationship identification system according to claim 24, wherein the control unit captures at least one feature image in the image region, and then selects one of each of the feature images. Performing a calculation on the first eigenvalue or the second eigenvalue corresponding to each pixel of the plurality of known sample images to obtain a similarity value of the plurality of known sample images respectively. And integrating a plurality of similarity values generated by comparing the feature image with the plurality of known sample images and sorting the plurality of similarity values to output a plurality of possible comparison comparison results. For example, the dynamic real-time relative relationship identification system described in claim 25, wherein the calculus may be a formal correlation comparison method or an approximation comparison method. The dynamic instant relative relationship identification system of claim 25, wherein each known sample image corresponds to a number or a character image. The dynamic instant relative relationship identification system according to claim 25, wherein the first feature value and the second feature value are respectively a combination of a weight value and a gray value, and the third feature value is a gray level. value. The dynamic instant relative relationship identification system of claim 22, wherein the image feature is the size, area or shape of the feature. The dynamic instant relative relationship identification system of claim 22, wherein the detected image feature is a size, an area or a shape of the feature on the movable device. The dynamic real-time relative relationship identification system described in claim 22 is disposed on a mobile vehicle. For example, the dynamic instant relative relationship identification system described in claim 22 further includes a second image capturing device that further captures the traffic signal state and the speed limit number image. The dynamic real-time relative relationship identification system according to claim 32, wherein the control unit determines whether the light is red or the speed of the mobile vehicle is close to or based on the traffic signal state and the speed limit image. When the speed limit is exceeded, a warning message is sent by a warning unit. The dynamic real-time relative relationship identification system according to claim 32, wherein the control unit determines, according to the traffic signal state and the speed limit number image, whether the light is red or the speed of the movable device is close to or exceeds At the time limit, the images captured by the first image capturing device and the second image capturing device are recorded.