KR102238124B1 - Detection System For Reverse Direction Moving Object - Google Patents

Abstract

The present invention relates to a system for detecting an object moving in reverse direction capable of allowing a control agent to intuitively determine the moving direction of the object by detecting the object based on a motion vector, displaying the same as complementary color visualization information according to the moving direction of the detected object, and providing the same to the control agent. The system for detecting the object moving in reverse direction comprises: an image packet receiver; a motion vector extractor; a color wheel module; and a mapping module.

Description

Detection System For Reverse Direction Moving Object}

The present invention relates to the technical field of analyzing an image and detecting a setting event for an object existing in an image.

Currently, technologies for calculating values through logical computational processes such as artificial intelligence (AI) and deep learning are rapidly developing. In addition, it is expanding to various technical fields, and various technologies connected in conjunction with the above-described technological development are also being developed. For example, a technology for calculating a specific value through a logical operation such as artificial intelligence is being combined with a closed circuit television (CCTV) to develop a selective control system technology that goes beyond the limitations of the past closed circuit television.

The screening task system technology is a technology that detects objects set in advance, provides information on movement of detected objects to monitoring personnel, and helps control personnel to monitor more effectively.

However, the currently developed screening task system is proceeding with a structure that converts images captured on most closed-circuit televisions into image packets, decodes the converted image packets into image frames, and detects abnormal objects among the decoded image frames. That is, the current screening task system decodes the received data into an image and analyzes it based on this in order to detect an accurate object.

The sorting task system of this type has a problem of using a lot of resources to detect an object and a problem in that it cannot accurately detect the type and location information of an object despite the use of a large number of resources. In particular, there is a problem in that it is not possible to distinguish between a vehicle traveling in a steady manner and a vehicle traveling in reverse, and an image capable of intuitively identifying a vehicle moving in different directions cannot be calculated.

Republic of Korea Patent Registration No. 10-2042397 (announcement date: 2019.11.8)

Accordingly, the present invention accurately detects an object based on a motion vector, and provides visualization information according to the movement direction of the detected object to the control personnel, thereby providing the control personnel intuitively to the problem of the existing system, that is, the driving direction of the object. We are trying to solve the problem of not providing data that can be grasped.

The problem to be solved of the present invention is not limited to the problems mentioned above, and other technical problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

The reverse-driving object detection system of the present invention for achieving the problem to be solved is connected to a camera unit that converts an image frame into an image packet in a data communication manner, and receives the image packet sequentially as time passes. An image packet receiver configured to receive and sequentially store the image packets, including a module and a storage module sequentially storing the image packets;

Including a parsing module for parsing the video packet by the video packet receiving unit and an extraction module for extracting a pixel block including a plurality of motion vectors and a plurality of motion vectors one by one, and the motion vector and A motion vector extractor for extracting a pixel block;

When the color wheel module and the motion vector are received, the color wheel module including the RGB color wheel having colors that are located opposite to each other at a position opposite to the center point of one wheel, and whose color is gradually darkened from the center point to the outside, is received. By overlapping each of the motion vectors on the RGB color wheel, the color of the overlapping area is extracted and mapped to a plurality of pixel blocks to form the pixel block into an RGB dynamic object having a plurality of colors and an RGB background object having a color. Including a mapping module that forms a single RGB color map,

A color map that generates an RGB dynamic object having a first color by a one-way motion vector and a complementary dynamic object having a second color that is a complementary color of the first color by a reverse motion vector opposite to the one-way motion vector. Generation unit;

A normalization module that sets any one of the color channel (H), saturation channel (S), and brightness channel (V) of the RGB color map as a reference channel, and converts the RGB color map to HSV color to convert the HSV color map. An HSV color conversion unit for converting the RGB color map into the HSV color map including an HSV color conversion module to form; And

A saturation value extraction module that extracts the saturation value of the HSV dynamic object of the HSV color map and the saturation value of the HSV background object of the HSV color map, and a median filter that removes noise from the image using a nonlinear near operation After rearranging the saturation value of the object and the saturation value of the background object, a filter module is provided to extract the intermediate density object by placing the intermediate saturation value in the center of the dynamic object. It may include a dynamic object detection unit for classifying background objects.

It may include a moving object output unit that receives the intermediate density object and the RGB background object, and outputs the intermediate density object and the RGB background object by overlapping them.

A fuzzy receiving module that receives the saturation and hue values of the HSV dynamic object and the HSV background object, and a reference fuzzy value and a reference fuzzy equation are set in advance, and the saturation of the HSV dynamic object and the HSV background object in the reference fuzzy equation A fuzzy value calculation module that calculates a first computational fuzzy value by putting a value and a color value of the HSV dynamic object and the HSV background object, and extracting a color corresponding to the first computational fuzzy value from the HSV color map to form a fuzzy map. Fuzzy map module,

Comparing the first computational fuzzy value and the reference fuzzy value, extracting only the first computational fuzzy value corresponding to the HSV dynamic object as a fuzzy dynamic object, and deleting the first computational fuzzy value corresponding to the HSV background object. The dynamic object display module may be provided, and may further include a fuzzy map generator configured to extract a fuzzy dynamic object from the fuzzy map by converting the HSV color map into a fuzzy map.

When the intermediate density object and the fuzzy dynamic object are received, a reverse running object discrimination unit for determining the intermediate density object and the fuzzy dynamic object as a reverse running object may be further included.

The reverse-running object detection system according to the present invention detects an object based on a motion vector, and displays it as visualization information of complementary colors according to the traveling direction of the detected object. And by providing this to the control personnel, the control personnel can intuitively determine the driving direction of the object.

1 is a block diagram of a system for detecting a reverse driving object according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a structure of an image frame photographed by the camera of FIG. 1 and a bitstream of the image frame.
3 is a diagram showing a state in which a color is extracted from the RGG color wheel by superimposing a one-way motion vector of an object included in an image frame and extracted on an RGB color wheel.
4 is a diagram showing colors filled in a dynamic object and a background object according to an operation of the color map generator of FIG. 3.
5 is a diagram showing a state in which a color is extracted from the RGG color wheel by superimposing a reverse motion vector of an object included in an image frame and extracted on the RGB color wheel.
6 is a view showing colors filled in a dynamic object and a background object according to the operation of the color map generator of FIG. 5.
7 and 8 are views showing an operating state of an HSV color conversion unit included in the system for detecting a reverse-driving object according to an embodiment of the present invention.
9 is a view showing an operation state of a fuzzy map generator included in the system for detecting a reverse driving object according to an embodiment of the present invention.
10 is a view showing an operating state of a reverse traveling object discrimination unit included in the reverse traveling object detection system according to an embodiment of the present invention.

Advantages and features of the present invention, and methods for achieving them will become apparent with reference to the embodiments described later in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms. It is provided to fully inform the person of the scope of the invention.

The claims of the present invention may be defined by the claims as well as the description supporting the claims. In addition, the same reference numerals refer to the same elements throughout the specification.

The reverse running object detection system described throughout this specification may be a computer that transmits and receives data and processes data.

First, a system for detecting a reverse-driving object according to an embodiment of the present invention will be generally described with reference to FIG. 1.

The reverse-driving object detection system 1 according to an embodiment of the present invention extracts a motion vector B generated in an image frame, and detects an object based on the extracted motion vector. In addition, it is presented as visualized information in complementary colors according to the driving direction of the detected object, and it is provided to the control personnel.

The present invention allows the control personnel to intuitively determine the driving direction of the object. For example, in the present invention, when the object travels in one direction, that is, in the forward direction, the dynamic object is represented by a first color, and when the dynamic object travels in the other direction, that is, the reverse direction, the dynamic object is represented by a second color that is the complementary color of the first color. In addition, the background color is represented by the first color so that the dynamic objects can be clearly distinguished. In addition, this image is provided to the control personnel, and the control personnel can intuitively grasp the driving direction of the dynamic object.

The reverse-driving object detection system 1 having such a characteristic includes an image packet receiving unit 10, a motion vector extracting unit 20, and a color map generating unit 30 as constituent elements. In addition, the reverse running object detection system 1 includes an HSV color conversion unit 40, a dynamic object detection unit 50, a moving object output unit 60, a fuzzy map generation unit 70, and a reverse running object discrimination unit 80. It may further include as a component.

Hereinafter, components constituting the reverse-driving object detection system according to an embodiment of the present invention and operations of each component will be described in detail with reference to FIGS. 1 to 10.

However, among the components of the present invention, the image packet receiving unit 10 and the motion vector extracting unit ( 20), a description of the color map generation unit 30 and a process of detecting the reverse movement of the object through the description will be described in detail. Thereafter, based on the description described above with reference to FIGS. 7 to 10, the HSV color conversion unit 40, the dynamic object detection unit 50, the moving object output unit 60, the fuzzy map generation unit 70, and the reverse running object A description of the determination unit 80 and a process through which the reverse driving of the dynamic object can be more accurately detected will be described.

Hereinafter, components included in the reverse running object detection system 1 will be described in detail.

The image packet receiving unit 10 is connected by wire or wirelessly to enable data communication with the camera unit A. In addition, it includes a receiving module 110 for receiving data transmitted from the camera unit A, that is, an image packet, and a storage module 120 for sequentially storing the same. Here, the camera unit A is formed of a closed circuit television (CCTV) to be a device for photographing a specific area. Such a camera unit A may generate an image frame 101 by capturing a certain area as shown in FIG. 2A including the generation module A1. In addition, the camera unit A may include the conversion module A2 to generate the image frame 101 as an image packet 102 as shown in FIG. 2B. In this case, the image packet 102 may be a bitstream including a header, a coding structure, a motion vector, and a plurality of macroblocks including image information.

The image packet receiving unit 10 may store the image packet 102 after receiving the image packet 102 from the camera unit A. For example, the image packet receiving unit 10 may sequentially receive the image packets from the camera unit A and then store them. In addition, the image packet receiving unit 30 may sequentially transmit the stored image packet 102 to the motion vector extracting unit 20.

The motion vector extraction unit 20 includes a parsing module 210 and an extraction module 220. Here, the parsing module 210 may parse the image packet. In addition, the extraction module 220 may extract a pixel block C including a plurality of motion vectors and a plurality of motion vectors B one by one.

The motion vector extracting unit 20 uses the parsing module 210 and the extraction module 220 to provide a motion vector B existing in each coding structure of the image packet 102 and a pixel block C including one motion vector. Can be extracted accurately. For example, the motion vector extractor 20 may extract a motion vector B and a pixel block C generated in the image frame 101 as shown in FIG. 3. Here, the extracted motion vector may be formed long when the image change is large, and may be formed short when the image change is small. In this case, when the object moves, the motion vector extracting unit 20 may extract a motion vector having characteristics of the shape and size of the moving object from the boundary between the objects by matching the coding structure of the moving object and the non-moving object.

Furthermore, the motion vector extractor 20 may set the extracted motion vector to 0 when the size of the extracted motion vector is smaller than the reference size, and may extract the extracted motion vector as a motion vector when it is larger than the reference size. Through this, the motion vector extractor 20 may extract a motion vector from the image packet.

However, in this specification, in order to help understanding the motion vector, a motion vector is shown as a reference for forming a motion vector in an image image.

In addition, the motion vector extracting unit 20 may further include a motion vector grouping module 230 to group the motion vectors B into figures having a predetermined size to form one motion vector group. For example, the motion vector grouping module 230 may form a motion vector group by grouping the motion vectors B into figures having a predetermined size, as shown in FIG. 3. As described above, the motion vector grouping module 230 may set a first object corresponding to a vehicle and a second object corresponding to a person as a setting object, and compare the setting object with the motion vector group. In addition, when the motion vector group corresponds to the first object, the motion vector group may be detected as one dynamic object, that is, a vehicle object. At this time, the motion vector grouping module 230 learns the motion vector group through Histogram of Oriented Gradients (HOG), Support Vector Machine (SVM), and deep learning, sets it as a first object, and is continuously updated. By contrasting the first object with the newly detected motion vector group, it is possible to more accurately detect the motion vector group as a dynamic object.

The motion vector extraction unit 20 accurately detects the dynamic object and transmits the motion vector in the detected object to the color map generation unit 30.

The color map generator 30 fills the pixel block C with any one of RGB colors. The color map generator 30 includes a color wheel module 310 including an RGB color wheel D and a mapping module 320 for forming an RGB color map. Here, the color wheel module 310 includes an RGB color wheel D having colors opposite to each other at positions opposite to each other with respect to the center point, and colors gradually becoming darker from the center point to the outside. Such an RGB color wheel D may be the same data as the RGB color wheel as shown in (b) of FIG. 3. When the motion vector B is received, the mapping module 320 superimposes each motion vector B on the RGB color wheel D shown in FIG. And a color similar to purple may be mapped to the plurality of pixel blocks C. In addition, the mapping module 320 is an example of a second color that becomes a complementary color of the first color by a reverse motion vector (RB) in a direction opposite to the one-way motion vector (B) as shown in FIG. 5A, The apricot color and a color similar to the apricot color may be mapped to the plurality of pixel blocks C.

The color map generation unit 30 includes an RGB dynamic object E1 having a color similar to purple and purple in the pixel block C of the dynamic object and an RGB background object having white ( It is formed by E2) and one RGB color map (E) can be formed. In addition, the color map generation unit 30 may further include an RGB binarization module 340, a background object conversion module 350, and a dynamic object application module 360. Here, the RGB binarization module 340 may set the reference brightness value to compare the brightness values of the dynamic object and the background object with the reference brightness value, and display the RGB dynamic object and the RGB background object in black or white. In addition, the background object conversion module 350 may convert the binary background object F2 expressed in black or white in the RGB binarization module 340 into any one color mapped to the pixel block C. In addition, the dynamic object application module 360 may apply the RGB dynamic object E1 formed in the RGB color wheel D to the binary dynamic object F1 expressed in white or black in the background object conversion module 350. .

The color map generation unit 30 compares the brightness values of the RGB dynamic object E1 and the RGB background object E2 to the reference brightness value in FIG. 4A through the RGB binarization module 340. Through this, as shown in (b) of FIG. 4, the RGB dynamic object E1 can be represented as a white binary dynamic object F1, and the RGB background object E2 can be represented as a black binary background object F2. . In addition, the color map generation unit 30 uses the background object conversion module 350 to map the binary background object F2 to the pixel block C as shown in FIG. 4C. Can be converted to color. And, using the dynamic object application module 360, the converted background object (G1) converted in the background object conversion module 350, the RGB dynamic object (E1) by a one-way motion vector (B) formed by the RGB color wheel (D). ) Can be applied.

In addition, the color map generator 30 may apply an RGB dynamic object based on a reverse model vector RB formed in the RGB color wheel D to the transition background object G1. As an example, as shown in FIG. 6, a second color that becomes a complementary color of the first color, that is, an apricot color and a complementary color dynamic object E3 having a color similar to the apricot color may be applied.

As described above, the present invention represents the RGB dynamic object E1 as the first color when the dynamic object travels in the forward direction through the image packet receiving unit 10, the motion vector extracting unit 20, and the color map generating unit 30, When the RGB dynamic object E1 travels in the reverse direction, a complementary color dynamic object E3 having a first color and a second color that is a complementary color may be displayed. In the present invention, when the dynamic object travels in the reverse direction, the color when the dynamic object travels in the background is indicated by the background color, and the dynamic object running in the reverse direction is indicated by the complementary color of the background color, so that the dynamic object running in the reverse direction can be clearly distinguished from the background. In addition, this image is provided to the control personnel, and the control personnel can intuitively grasp the driving direction of the dynamic object.

Hereinafter, based on the description described above with reference to FIGS. 7 to 10, the HSV color conversion unit 40, the dynamic object detection unit 50, the moving object output unit 60, the fuzzy map generation unit 70, and the reverse running object A description of the determination unit 80 and a process in which the reverse driving of the dynamic object is more accurately detected through this will be described in detail.

The HSV color conversion unit 40 converts the RGB color map into an HSV color map. The HSV color conversion unit 40 includes a normalization module 410 and a color conversion module 420. Here, the normalization module 410 sets any one of a color channel, a saturation channel, and a brightness channel of the RGB color map as a reference channel. And the HSV color conversion module 420 converts the RGB color map to HSV color.

The HSV color conversion unit 40 converts the data of the RGB color model shown in FIG. 7 into data of the HSV color model using the following [Equation 1], [Equation 2], and [Equation 3].

[Equation 1]

[Equation 2]

[Equation 3]

Here, among the data of the RGB color model, R stands for Red, G stands for Green, and B stands for Blue. In addition, among the data of the HSV color model, H stands for hue, S stands for saturation, and V stands for brightness.

The data of the HSV color model includes color data (H), saturation data (S), and brightness data (V). The HSV color model constructs colors using a color channel (H), a saturation channel (S), and a brightness channel (V) based on human intuitive vision. In the case of such an HSV color space, since the brightness (V) portion can be separated, it is possible to more effectively grasp the amount of light than in the case of the RGB color space.

The HSV color conversion unit 40 having these characteristics sets the channel for the amount of light, that is, the brightness channel as a reference channel, and converts the RGB color map as shown in FIG. 8A to the HSV color as shown in FIG. 8B. Convert to map (H).

The dynamic object detection unit 50 can more clearly distinguish between the HSV dynamic object H1 and the HSV background object H2 of the HSV color map H, including the saturation value extraction module 510 and the filter module 520. have. Here, the saturation value extraction module 510 may extract a saturation value of a dynamic object and a saturation value of a background object of the HSV color map. In addition, the filter module 520 rearranges the saturation value of the dynamic object and the saturation value of the background object by generating a median filter that removes noise from the image by using a nonlinear near operation. Thereafter, the intermediate saturation value can be placed in the center of the dynamic object. For example, the dynamic object detection unit 50 has a saturation value of the HSV dynamic object H1 and a saturation value of the HSV background object H2 as shown in FIG. 8C. 102, 67, 84, 224, 189 , 33, 212, 163, 54, after rearranging them into 33-54-67-84-102-163-189-212-224 in small order, 102 can be placed in the center of the HSV dynamic object (H1). In addition, the dynamic object detection unit 50 may extract the intermediate concentration object I1 in which the intermediate saturation value is located through the filter module 520.

The dynamic object detection unit 50 more clearly distinguishes the HSV dynamic object H1 and the HSV background object H2, extracts the intermediate concentration object I1, and transmits it to the moving object output unit 60.

The moving object output unit 60 receives the intermediate density object (I1) from the dynamic object detection unit 50, receives the RGB background object (E2) of the RGB color map from the color map generator 30, and then overlaps it. You can print it out as one image.

The fuzzy map generator 70 may convert the HSV color map H into a fuzzy map J as shown in FIG. 9A to extract the fuzzy dynamic object J1 from the fuzzy map J. The fuzzy map generation unit 70 includes a fuzzy reception module 710, a fuzzy value calculation module 720, a fuzzy map module 730, a dynamic object display module 740, and a fuzzy binarization module 750. Here, the fuzzy receiving module 710 may receive saturation values and color values of the HSV dynamic object H1 and the HSV background object H2 of the HSV color map H. In addition, the fuzzy value calculation module 720 has an input color value, a reference fuzzy value, and a reference fuzzy equation set in advance, and the reference fuzzy equation is calculated by substituting the saturation value of the dynamic object and the background object and the color value of the dynamic object and the background object. One operation purge value can be calculated. Further, the fuzzy map module 730 may form a fuzzy map H by extracting a color corresponding to the first calculated purge value from the HSV color map H. In addition, the dynamic object display module 740 may extract a first computational fuzzy value corresponding to the HSV dynamic object as a fuzzy dynamic object J1 by comparing the first computational fuzzy value and the reference fuzzy value. In addition, the remaining first computational fuzzy values can be deleted. Further, the fuzzy binarization module 750 may binarize the fuzzy dynamic object J1 and the deleted object, that is, the fuzzy background object.

As an example, the fuzzy map generation unit 70 calculates a first computational fuzzy value using the fuzzy value calculation module 720 using the HSV color map H shown in FIG. 9A using the following equation. . In addition, by using the fuzzy map module 730 to extract a color corresponding to the first computational fuzzy value from the HSV color map H, the fuzzy map J of FIG. 9B may be generated.

[Equation 4]

[Equation 5]

In [Equation 4], μ _c denotes a fuzzy function, a denotes a constant, h _c denotes a color of interest value, h denotes an input color value, and W and H denote the image height and Is the width.

In [Equation 5], a curly fuzzy map μ _H which is made of the vehicle's normal driving direction color fuzzy set μ _c1 and the reverse driving direction color fuzzy set μ _{c2 as maximum values of the vehicle generated using [Equation 4].} to be. In addition, the fuzzy map generation unit 70 uses the dynamic object display module 740 to compare the first computational fuzzy set and the reference fuzzy value, as shown in Fig. 9(b), the first computational fuzzy value corresponding to the dynamic object. Can be extracted as a fuzzy dynamic object (J1). In addition, the first computational fuzzy value corresponding to the background object can be deleted. In addition, the fuzzy map generator 70 may use the fuzzy binarization module 750 to represent the fuzzy dynamic object J1 in white and the fuzzy background object J2 in black as shown in FIG. 9C.

As shown in FIG. 10, when the reverse running object discrimination unit 80 receives both the intermediate concentration object I1 and the fuzzy dynamic object J1 extracted by the fuzzy binarization module 750, the received object is converted to the reverse running object (K). ).

The reverse-driving object detection system 1 is composed of an image packet receiving unit 10, a motion vector extracting unit 20, and a color map generation unit 30, HSV color conversion unit 40, and dynamic object detection unit 50. , The moving object output unit 60, the fuzzy map generation unit 70, and the reverse running object discrimination unit 80 may be included as components to more clearly represent a dynamic object running in reverse. And by providing such an image to the control personnel, the control personnel can more accurately determine the reverse running object.

Embodiments of the present invention have been described above with reference to the accompanying drawings, but those of ordinary skill in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features. You will be able to understand. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not limiting.

1: reverse-driving object detection system 10: video packet receiver
101: image frame 102: image packet
110: receiving module 120: storage module
20: motion vector extraction unit 210: parsing module
220: extraction module 230: motion vector grouping module
30: color map generation unit 310: color wheel module
320: mapping module 330: motion vector grouping module
340: binarization module 350: background object conversion module
360: dynamic object application module 40: HSV color conversion unit
410: normalization module 420: HSV color conversion module
50: dynamic object detection unit 510: saturation value extraction module
520: filter module 60: moving object output unit
70: fuzzy map generation unit 710: receiving module
720: fuzzy value calculation module 730: fuzzy map module
740: dynamic object display module 80: reverse driving object discrimination unit
A: Camera part A1: Generation module
A2: Conversion module B: Motion vector
RB: Reverse motion vector C: Pixel block
D: RGB color wheel E: RGB color map
E1: RGB dynamic object E2: RGB background object
E3: Complementary dynamic object F1: Binary dynamic object
F2: Binaryized background object G1: Transition background object
H: HSV colormap H1: HSV dynamic object
H2: HSV background object I1: Medium concentration object
J: Fuzzy map J1: Fuzzy dynamic object
K: reverse running object

Claims (5)

delete A receiving module 110 connected to the camera unit A for converting an image frame into an image packet and capable of data communication, and sequentially receiving the image packet over time, and a storage module sequentially storing the image packet An image packet receiver 10 that receives and sequentially stores the image packets, including 120;
A parsing module 210 for parsing the image packet by the image packet receiver 10 and an extraction module for extracting a pixel block C including one of a plurality of motion vectors and a plurality of the motion vectors B Including 220, a motion vector extracting unit 20 for extracting the motion vector and the pixel block from the video packet;
A color wheel module 310 including an RGB color wheel D having colors that are opposite to each other at positions opposite to each other with respect to the center point of one wheel, and the color gradually darkens from the center point to the outside,
When the motion vector (B) is received, each of the motion vectors (B) is superimposed on the RGB color wheel (D) to extract the color of the overlapping area and map it to a plurality of pixel blocks (C), and the pixel block (C) is ) To form an RGB dynamic object (E1) having a plurality of colors and an RGB background object (E2) having one color to form a single RGB color map (E),
An RGB dynamic object having a first color is generated by a one-way motion vector, and a complementary dynamic object (E3) having a second color that is a complementary color of the first color by means of a reverse motion vector (RB) opposite to the one-way motion vector. Including a color map generating unit 30 to generate ),
A normalization module 410 that sets any one of the color channel (H), saturation channel (S), and brightness channel (V) of the RGB color map as a reference channel, and the RGB color map (E) as HSV color. An HSV color conversion unit 40 for converting the RGB color map H to the HSV color map H, including an HSV color conversion module 420 for converting to form an HSV color map H;
Image noise is reduced by using a saturation value extraction module 510 for extracting the saturation value of the HSV dynamic object (H1) of the HSV color map and the saturation value of the HSV background object (H2) of the HSV color map and a nonlinear near operation. A filter module 520 that generates a median filter to be removed and rearranges the saturation value of the dynamic object and the saturation value of the background object, and then locates the intermediate saturation value at the center of the dynamic object to extract the intermediate density object. The system further comprises a dynamic object detection unit (50) that is provided and separates the HSV dynamic object (H1) and the HSV background object (H2) of the HSV color map. The method of claim 2,
Receiving the intermediate density object (I1) and the RGB background object (E2), including a moving object output unit 60 for outputting the intermediate density object (I1) and the RGB background object (H2) overlapping, Reverse running object detection system. The method of claim 3,
A fuzzy receiving module 710 that receives the saturation and color values of the HSV dynamic object (H1) and the HSV background object (H2), and a reference fuzzy value and a reference fuzzy equation are preset, and the HSV A fuzzy value calculation module 720 for calculating a first computational fuzzy value by putting the saturation value of the dynamic object and the HSV background object and the color value of the HSV dynamic object and the HSV background object, and the HSV color map H A fuzzy map module 730 for forming a fuzzy map by extracting a color corresponding to the first calculated fuzzy value,
Comparing the first computational fuzzy value and the reference fuzzy value, only the first computational fuzzy value corresponding to the HSV dynamic object is extracted as a fuzzy dynamic object (J1), and the first computational fuzzy value corresponding to the HSV background object (H2) A dynamic object expression module 740 that deletes the computational fuzzy value is provided, and a fuzzy map is generated to extract the fuzzy dynamic object (J1) from the fuzzy map (J) by converting the HSV color map (H) into a fuzzy map (J). Further comprising a unit 70, the reverse running object detection system. The method of claim 4,
When the intermediate concentration object (I1) and the fuzzy dynamic object (J1) are received, a reverse running object discrimination unit 80 that determines the intermediate concentration object (I1) and the fuzzy dynamic object (J1) as a reverse running object (K) Further comprising, a reverse running object detection system.

Images