KR102189063B1 - Apparatus, system and method for determining abnormality on a road - Google Patents

Abstract

According to the present invention, provided is an image analysis device for recognizing an abnormal situation on the road. The image analysis device may comprise: a block mapping unit for mapping a road detected in a road image into blocks; an object detection unit detecting a block in which an object enters or exits from the blocks; a traffic state calculation unit for calculating a traffic state value of a block adjacent to the detected block in accordance with the entry or exit of the object in the detected block; and an event generation unit for comparing the traffic state value with a preset value, and determining that an event has occurred for the detected block when the comparison result satisfies a predetermined condition.

Description

Apparatus, system and method for determining abnormality on a road}

The present invention relates to an apparatus, system, and method for determining an abnormal situation, and more particularly, to detect the state of the flow through object recognition and tracking in an image captured of the movement (progress) flow of a moving object, and The present invention relates to recognizing a change in lane and determining whether there is an abnormal situation on a road or passage displayed in the image accordingly.

Conventionally, a smart (intelligent) closed circuit television (CCTV) for detecting an unexpected situation (abnormal situation) by photographing a traffic flow is difficult to judge an accident situation, and there are excessively many problems in the occurrence of an event for a stopped vehicle. For example, in the case of a road in which a designated section such as commuting time or holidays is formed long, a designated body occurs as the traffic volume increases. Even in this case, if an event for all stopped vehicles occurs, confirmation is not required. In spite of the occurrence of the event, the primary purpose of detecting an unexpected or abnormal situation on the road cannot be achieved, and there is an inefficient aspect of the resource use of the detection system.

The present invention is to improve the device, system, or method for detecting an unexpected or abnormal situation described above, and in more detail, to propose an apparatus, system, or method capable of detecting an unexpected or abnormal situation, which will be described later. We want to achieve it through solutions.

The problems to be solved to be achieved in the present invention are not limited to the problems to be solved, and other problems that are not mentioned can be clearly understood by those of ordinary skill in the technical field to which the present invention belongs from the following description. There will be.

According to the present invention, an image analysis device for recognizing an abnormal situation is proposed, wherein the image analysis device includes: a block mapping unit for mapping a road detected in a road image into blocks; An object detection unit detecting a block in which an object enters or exits from among the blocks; A traffic state calculator configured to calculate a traffic state value of a block adjacent to the detected block according to the entry or exit of the object in the detected block; And an event generator configured to compare the traffic state value with a preset value and, if the comparison result satisfies a predetermined condition, determine that an event has occurred for a block that satisfies the predetermined condition.

Additionally or alternatively, the device may include the traffic state value and a weight of the block.

Additionally or alternatively, when the object is detected to have advanced from the detected block, a weight applied to a block preceding the detected block may be determined.

Additionally or alternatively, when it is detected that the object has entered the detected block, a weight applied to a block after the detected block may be determined.

Additionally or alternatively, when the number of objects advancing from the detected block is greater than or equal to a threshold, a weight applied to a block preceding the detected block may be determined.

Additionally or alternatively, when the number of objects entering the detected block is greater than or equal to a threshold, a weight applied to a block after the detected block may be determined.

Additionally or alternatively, if the traffic state value for all blocks is less than a preset value, the device may detect the occurrence of an event by calculating traffic state values for blocks in the road image.

Additionally or alternatively, the traffic state value is based on at least one of an average moving speed of objects passing through a block or an average stopping time of objects in a block, and may be obtained through a radar sensor or image processing.

Additionally or alternatively, when there are a plurality of the mapped blocks, the object detection unit, the traffic state calculation unit, and the event generation unit may be processed for each block in order of magnitude of the traffic state value of each block.

Additionally or alternatively, for blocks with the same traffic state value, a block close to the block with a larger traffic state value around it may be processed first.

Additionally or alternatively, the device may further include an image acquisition unit that acquires the road image.

Additionally or alternatively, the object detection unit may further include a radar sensor for detecting and tracking the object.

Additionally or alternatively, the device may further include an operating unit that receives the image and stores or displays it or receives and displays information on the occurrence of the event.

Additionally or alternatively, the traffic state value may include one of a plurality of values according to a section to which the average moving speed of the object in each block belongs.

Additionally or alternatively, the traffic state value may include one of a plurality of values according to a section to which the average stop time of the object in each block belongs.

Additionally or alternatively, the device may further include an operation unit receiving information on an object from the object detection unit and generating and displaying at least one of traffic information, illegal parking information, speed violation information, and life crime prevention information.

In accordance with the present invention, a method for recognizing an abnormal situation is proposed, which is performed by an image analysis apparatus, comprising: mapping a road detected in a road image to blocks; Detecting a block in which an object enters or exits from among the blocks;

Calculating a traffic state value of a neighboring block of the detected block according to the entry or exit of the object in the detected block; And comparing the traffic state value with a preset value, and if the comparison result satisfies a predetermined condition, determining that an event has occurred for a block that satisfies the predetermined condition.

The above problem solving methods are only some of the embodiments of the present invention, and various embodiments in which the technical features of the present invention are reflected are based on the detailed description of the present invention to be described below by those of ordinary skill in the art. Can be derived and understood.

According to the present invention, it is possible to accurately and efficiently detect an unexpected situation or an abnormal situation of a moving flow of a moving object on a road or passageway.

In addition, according to the present invention, it is possible to effectively reduce the occurrence of a stop vehicle event that is unnecessary even when a road or passage is congested.

In addition, according to the present invention, since an event occurrence (indicated as an unexpected block) is displayed in block units, the occurrence of an abnormal situation is displayed in a block form in a control center, etc., so that an operator can intuitively check whether an abnormal situation has occurred and its location.

The effects that can be obtained in the present invention are not limited to the effects mentioned above, and other effects not mentioned can be clearly understood by those of ordinary skill in the art from the following description. will be.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included as part of the detailed description to aid in understanding of the present invention, provide embodiments of the present invention, and together with the detailed description will be described the technical idea of the present invention.
1 shows the overall configuration of a traffic control system related to the present invention.
2 and 3 are examples of recognizing a vehicle on a road according to the present invention and processing the image.
4 illustrates an evolution process of congestion over time in a form in which a road or passage according to the present invention is blocked.
5 is a block diagram showing a part of a road or passage.
6 shows a flowchart of a method for recognizing an abnormal situation according to the present invention.
7 shows a configuration of an apparatus for recognizing an abnormal situation according to the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described in the present specification, and may be implemented in various different forms. The terms used in the present specification are intended to aid understanding of the embodiments, and are not intended to limit the scope of the present invention. In addition, the singular forms used below also include plural forms unless the phrases clearly indicate the opposite meaning.

1 shows the overall configuration of a traffic control system in the field to which the present invention belongs. The system includes an image acquisition device (1) such as a camera for capturing an image by photographing a traffic situation, and accessories related thereto, a server (2) connected to the image acquisition device by wire or wirelessly, and the image It may include a control system (3) such as a traffic control room operated by various local governments, road works, facility corporations, and the National Police Agency, which perform tasks such as real-time control or storing and managing images. The image acquisition device may operate in conjunction with a vehicle detection means such as a vehicle detection sensor (eg, at least one of an electromagnetic sensor, a radar sensor, and an image sensor). The apparatus, system, etc. according to the present invention may be included in any one of the configurations shown in FIG. 1 or may exist as a separate apparatus, system, or the like.

2 is an example of recognizing a vehicle according to the present invention and processing the image. 2 is a description based on an image taken of a road dedicated to a vehicle, but the present invention is not necessarily limited to a vehicle (or vehicle), and is applied to all moving objects (eg, vehicles, hereinafter simply referred to as “objects”). It is noted in advance that the present invention can be applied.

As shown in FIG. 2, it is possible to designate a lane in which vehicles are traveling, and to detect a (traffic) traffic state using the moving speed of each lane, and this may be used to determine abnormal situations such as stop, accident, traffic abnormal situation, etc. . The measurement of the moving speed on the lane may be measured through a device such as an electromagnetic sensor, a radar sensor, or an image sensor, or through image processing. For example, the traffic state may be determined as "smooth" when the average moving speed is 50 km/h or more, "slow" when the average moving speed is 50 km/h or more, and "congestion" when it is less than 30 km/h. Each can also be expressed as a quantitative value. For example, "smooth" may have a value of 0.5, "slow" may have a value of 0.3, and "congestion" may have a value of 0.1, which may be referred to as a communication state value. In addition, the reference speed for determining the communication state is only an example and does not limit the scope of the present invention.

In addition, the traffic status can be detected using the time of the stopped vehicle on each lane, and this can be used to determine abnormal situations such as stops, accidents, and traffic abnormalities. For example, the communication state may be judged as "smooth" in the case of a stop within 15 seconds, "slow" in the case of a stop of 15 seconds to 30 seconds or less, and "congestion" in the case of a stop exceeding 30 seconds. And each of these can also be expressed as a quantitative value. For example, "smooth" may have a value of 0.5, "slow" may have a value of 0.3, and "congestion" may have a value of 0.1, which may be referred to as a communication state value. In addition, the reference stop time for determining the communication state is only an example and does not limit the scope of the present invention.

In summary, the communication state value may be defined to include one of a plurality of values according to a section to which the average moving speed of an object in each block belongs. In addition, the communication state value may be defined to include one of a plurality of values according to a section to which the average stop time of the object in each block belongs.

3 is an example of recognizing a vehicle according to the present invention and processing the image. The left side of FIG. 3 shows a state in which one vehicle is individually detected and tracked, and the right side is a bird-eye view or top view of the detected and tracked vehicle through an IPM (Inverse Perspective Mapping) algorithm. -view) was created. The mapping coordinates of the IPM are based on a preset lane area.

A block for determining a traffic flow may be configured by dividing a lane with the same length using the bird-eye-view and designated lane information. The width of the lane divided into one block is set to about 3 meters, and the length (height) can be set to 30 meters. That is, the width of one block may be set to correspond to one lane/passage line. When congestion occurs, it is assumed that there are 5 cars in one block based on small cars, and the state of multiple objects instead of a single object can be used. In addition, the block may be generated through virtual space mapping for both straight and curved roads.

When the lane is divided into blocks and represented by virtual space mapping, it may be as shown in FIG. 4.

4 illustrates the evolution process of congestion over time in the form of blockage of a road according to the present invention. In the following description, "roads" are described, but since the present invention can be applied to all roads through which objects can pass, the name "roads" does not limit the scope of the present invention. Accordingly, the name of the “lane” on the “road” also does not limit the scope of the present invention. Congestion is expected to have five evolutionary processes, and each of the five shown in FIG. 4 corresponds to a block for the same point.

4A shows an initial state, and in this case, it is assumed that the traffic state of the road is smooth. In case, an abnormal situation may occur within one block, and FIG. 4B shows this situation. Then, if an event such as a collision accident actually occurs in the block corresponding to "slowing" in Fig. 4(b), the slowing or congestion is spread to the blocks after that block, which corresponds to "slowing". It tries to identify by changing lanes in the blocks after the block. Arrows from one block in FIG. 4 to another block indicate lane changes. The change of the lane can also be expressed as a block change of the object. In this case, the change of the adjacent front and rear blocks based on the moving direction must be excluded from the lane change, and the change of the block in the reverse direction of the moving direction must also be excluded. do. That is, it must be a block change corresponding to the arrows shown in (b) and (c) of FIG. 5 for one block.

4(c) to 4(e) illustrate a situation in which congestion spreads.

When slowing or spreading of congestion occurs according to the time sequence shown in FIG. 4, blocks in a congested state in FIG. 4 (e) form a cluster, and there is a high possibility that an abnormal situation has occurred in the block in front of them. With this assumption, a method for detecting an abnormal situation will be described.

FIG. 5 is a block diagram of a part of the road described in FIG. 4 and shows nine blocks.

FIG. 5A shows the identification codes of each block in order to identify 9 blocks, indicating that the top direction is the moving direction. In the remaining Figs. 5B and 5C, the identification codes of the blocks are the same. Also, an arrow passing through a block means movement of an object, and only lane changes are indicated by arrows.

FIG. 5B illustrates a situation in which a lane change is made to a block 22 among nine blocks (entry situation). The lane change proceeds from the blocks 21 and 23 to the block 22, and in this case, the blocks 11 and 13 (hatched) in front of the block before the lane change, respectively, or the block after the block 22 ( 32) It can be predicted that there is a cause of congestion in (hatched). Accordingly, a weight for calculating a traffic state value may be applied to the blocks 11 and 13, which will be described in more detail with respect to the flowchart of FIG. 6.

In addition, the lane change proceeds from the blocks 31 and 33 to the block 22, and in this case, it can be predicted that there is a cause of congestion in the rear block 32 (hatched) of the entry block 22. Accordingly, a weight for calculating a traffic state value may be applied to the block 32, which will be described in more detail with respect to the flowchart of FIG. 6.

In addition, FIG. 5C illustrates a situation in which a lane change is made from the middle block 22 to the other blocks 11, 21, 13, and 23 among the nine blocks (advancing situation). In this case, it can be predicted that there is a cause of congestion in the block in front of the block 22 before the lane change. Accordingly, a weight for calculating a traffic state value may be applied to the block 12 (marked with hatching), which will be described in more detail with respect to the flowchart of FIG. 6.

Meanwhile, the application of the weight may be set to be allowed only when the communication state of the block to which the weight is to be applied is "slow" or "congested". For example, referring to (d) or (e) of FIG. 4, a block in front or above the frontmost (or top) "congestion" block (hereinafter, "abrupt/abnormal situation block") is in a "smooth" state. This may be because, after passing through the delay or congestion caused by the “abrupt/abnormal situation block”, all lanes are in a “smooth” state, so it may be because it is a lane/line change regardless of the congestion or delay of the corresponding blocks.

For example, if an object in the "abrupt/abnormal situation block" moves to the left or right block of the block due to an accident in front of the same block, according to the above description, the "abrupt/abnormal situation block" Weights must be applied to the block immediately preceding it. However, since the block immediately preceding the "abrupt/abnormal situation block" will actually have a "smooth" communication state, it may not be necessary to apply a weight in calculating a traffic state value for detecting the abnormal situation.

Therefore, referring to (b) and (c) of FIG. 5, when determining the lane/passage change for the block 22, the blocks to which the weight may be applied are 11, 12, 13, 21, 23, 32, and only when at least one of the blocks 11, 12, 13, 21, 23, and 32 is in a delayed or congested state, it may be set to determine a lane/path change for the corresponding block 22. Contrary to this consideration, it is possible to set the above-described lane/passage change to be determined only for the neighboring blocks on the left and right sides of the “blocks in a delayed or congested state” and the immediately following blocks, and the neighboring blocks on the left and right of the rear block.

Also, since the "block" is a virtual area covering a certain area, there are several objects and can be moved. Thus, for example, a change in lane/line of traffic of one or two objects may be arbitrary movements of those objects irrespective of the traffic state (ground or congestion). Therefore, in order to exclude such arbitrary movement from calculating the traffic state value or calculating the weight for calculating the traffic state value, only when the number of lane/passage changes is equal to or greater than (or exceeding) a threshold value, the "change" for weight application is used. You need to be aware.

For example, referring to (c) of FIG. 5, only when the lane/passage change from block 22 to block 11, 13, 21, and 23 is N times (N is a natural number) or more (i.e., lane/line Only when the number of changed objects is N or more), weight application for block 12 may be allowed. Similarly, referring to (b) of FIG. 5, only when the lane/passage change is N or more times from the specific block 21 to the 22nd block (that is, only when the number of objects that have changed lanes/passage lines is N or more) ), it is possible to allow the application of weights to the block 11. Of course, as described above, when reviewed based on block 22, only if the communication state of at least one block among blocks 11, 12, 13, 21, and 23 is "delayed" or "congested", block 22 is You can try to detect a change in lane/line of traffic. Referring to FIG. 5B, only when the lane/passage change from specific blocks 31 and 33 to block 22 is N or more times (that is, only when the number of objects that have changed lanes/passage lines is N or more) ), it is possible to allow the application of the weight to the block 32. Of course, as described above, when reviewed based on block 22, it is possible to attempt to detect a lane/line change for block 22 only when the traffic status of block 32 is “delayed” or “congested”. .

6 shows a flowchart of a method for recognizing an abnormal situation according to the present invention.

Although the subject of performing the procedure according to FIG. 6 is not specified, it may be performed by an image analysis device for determining an abnormal situation according to the present invention, or a specific component (eg, a control unit or a processor unit) in the image analysis device, The names or names of the above components do not limit the scope of the rights of the content.

The device may receive an image of a road (S611). Since this is for determining an abnormal situation, the image must be input simultaneously or immediately after the photographing. The device may detect an object from the input image (S612) and track the object (S613). In this case, a radar sensor or the like may be used to detect and track the object.

Then, the device may map all or part of the road into blocks (S614), and for block mapping, refer to FIGS. 3 to 5 and descriptions thereof.

The device may update the block state for each block or may initially assign a value for the block state (S615). The block state means the communication state of the corresponding block described above, and may be given as one of "smooth", "slow" or "congestion", and may be quantified and have a corresponding value (ie, a communication state value).

After that, the device may select one of the mapped blocks in S614 (S616). One is selected in order to perform S617 to S622, which will be described later, and in S623, it is determined whether lane/passage change confirmation and subsequent procedures for all blocks have been performed. That is, S617 to S622 to be described later may be sequentially performed for one block, and then S617 to S622 may be performed for the next block (if any).

Alternatively, the device may select one of the neighboring blocks (enclosing the block) around the block whose block state is "slow" or "congested" from among the blocks mapped in S614 (S616). Alternatively, in S615, the device has the left and right blocks of the block whose block status is "slow" or "congested" (when the progress direction is based on "forward") and the block immediately behind and left and right blocks of the rear block. One of them may be selected (S616). As described in connection with FIG. 5, in practice, weight application is meaningful to a “slow” or “congested” block, and accordingly designates a “candidate block” that can determine weight application to a “slow” or “congested” block. , One of them may be sequentially selected to perform S617 to S622, which will be described later. In this case, procedures such as S616 to S624 may be performed only for the left and right blocks of the block in which the block state is "slow" or "congested", the immediately following block and the left and right blocks of the rear block.

In addition, in S616, it is said that one block is selected from "all blocks mapped in S614" or "(specific) neighboring blocks of blocks whose block state is a specific value", but when selecting one block, The block state (communication state) may be set to be preferentially selected from a poor block (ie, in the order of "congestion", "lag", and "smooth"). Further, for blocks having the same block state, it may be set so that a block close to a block having a poor block state may be preferentially selected. In the case of blocks located in the same block state or at the same distance, the order may be selected at random.

The device may detect whether an object lane/passage change has occurred for the selected block (S617). If a lane/passage change occurs, the device may detect whether the change of the lane/passage line of the object enters the selected block or advances from the selected block (S618). If no lane/line change has occurred, the procedure can proceed to S621.

When it is detected that the change of the lane/passage line of the object advances to the selected block, the device may increase or apply a weight to the block in front of the selected block (S619). When it is detected that the lane/passage change is entering the selected block, the device may increase or apply a weight to a block behind the selected block (S620).

Then, the device may calculate or update the traffic state value of the selected block (S621). The value of the traffic state may be determined by adding the weight of the corresponding block and the value of the traffic state. In this step, the device may calculate or update traffic state values for all blocks or the remaining "candidate blocks" as well as the selected block.

The device may compare the traffic state value of the selected block with a preset value. As an example of this comparison, the device may determine whether the traffic state value of the selected block exceeds a preset value (S622). In this step, the device may determine whether a traffic state value exceeds the preset value for all blocks or the remaining "candidate blocks" as well as the selected block.

When the comparison result satisfies a predetermined condition, the device may determine that an event has occurred for a block that satisfies the predetermined condition. As an example of this operation, the device may determine that an event has occurred for a block in which the traffic state value exceeds the preset value (S624). Then, the device may transmit information on the occurrence of the event to an embedded display unit or other system.

After receiving the information on the occurrence of the event, the display unit may display the occurrence of the event (indicated as an unexpected block) in block units, as shown in FIG. 4E. Accordingly, the occurrence of the abnormal situation is displayed in a block form in the control center, etc., so that the operator can intuitively check whether the abnormal situation has occurred and its location.

If the traffic state value for the block is less than the preset value, the device checks lane/line change for another block in the image (S617 to S620) and/or its subsequent procedure (S621 to S622, or S624). It may be checked whether or not (S623). If, in relation to S616, it is decided to perform the related procedure only for the "candidate block" rather than all the blocks in the image, the target blocks of S623 will correspond to other blocks among the "candidate blocks". If there is a block in the image that has not yet performed S617 and the subsequent procedure, the device may proceed to S616 again to select another block, and perform S617 and subsequent procedures.

If there are no blocks in the image in which S617 and the subsequent procedure have not been performed, the device returns to the image input step (S611), and may perform the subsequent procedure again on the image.

Meanwhile, when it is determined that the event has occurred (S624), traffic state values or weights for the blocks may be initialized.

In addition, in S616, the block status (communication status) obtained in S615 of the blocks was described as a reference when selecting the block, but alternatively, block selection is performed based on the traffic status value of each block calculated or updated in S621. It could be.

7 shows a configuration of an apparatus for recognizing an abnormal situation according to the present invention.

The image analysis apparatus includes a block mapping unit 71, an object detection unit 72, a traffic state calculation unit 73, an event generation unit 74, a control unit 75, a communication unit 76, and a storage unit 77. can do.

The block mapping unit 71 may map a road detected in a road image into blocks. The block may have a constant size, or may have a variable size, and the size of the block may be set smaller as the road farther away. The mapping of blocks has been described with reference to FIG. 4, and roads in the image are divided into unit blocks of a certain width and length (height).

The object detection unit 72 may detect or detect a lane change of an object in each of the blocks. Then, for each corresponding block in which the lane change of the object is detected, the object detection unit 72 may detect whether the object has entered the corresponding block from another block or has advanced from the corresponding block to another block. . To this end, the object detection unit may include a radar sensor for detecting and tracking the object.

Meanwhile, the object detection unit 72 may detect and recognize objects such as a person, vehicle, or vehicle license plate by classifying them. This is to be used as a purpose for various purposes. Specifically, the object detection unit 72 is a traffic information collection system, an illegal parking/stop control system, an unmanned road security system, a life security system, and intelligent It can operate in conjunction with CCTV systems classified as object tracking systems, unmanned signal/speed enforcement systems, and parking management systems. To this end, the object detection unit 72 detects and recognizes objects necessary for traffic information collection, illegal parking, speed violation or daily crime prevention, and provides information on the detected and recognized objects to various local governments, road works, facility corporations, It can be provided by a control system (or operation department) such as a traffic control room operated by the National Police Agency.

The traffic state calculator 73 may calculate or update a traffic state value for a neighboring block of the block according to the entry or exit of the object in the block. Here, the “peripheral block” refers to the descriptions related to FIGS. 5 and 6 described above. The traffic state value is a sum of the weight and the traffic state value for each block, and the traffic state corresponds to one of the traffic conditions of the corresponding block, such as "smooth", "delay" or "congestion", and In other words, the value of the communication status is quantified and expressed as a value. The traffic state value is based on at least one of a moving speed of objects moving or passing through a corresponding block or an average stopping time of objects in the corresponding block, and may be obtained through a radar sensor or image processing.

The object detection unit 72 may determine a weight applied to a specific block according to the entry or exit of the detected object. A block to which a weight to be described later is applied is determined according to whether the block change detection is "entering" or "going out". For the block to which the weight is applied, refer to FIG. 5.

That is, when it is detected that the object has advanced from the detected block, a weight applied to a block preceding the detected block may be determined. In addition, when it is detected that the object has entered the detected block, a weight applied to a block following the detected block may be determined.

Alternatively, when the number of objects advancing from the detected block is greater than or equal to a threshold, a weight applied to the block preceding the detected block may be determined, and when the number of objects entering the detected block is greater than or equal to the threshold, the detection A weight applied to a block after the block may be determined.

Here, the determination of the weight is a generic concept including an increase in the weight of the detected block.

The event generator 74 may compare the traffic state value of the block with a preset value, and if the comparison result satisfies a predetermined condition, it may determine that an event has occurred for a block that satisfies the predetermined condition. . For example, the event generator 74 determines whether the traffic state value of the block exceeds or is the same as a preset value, and if it exceeds or equals the preset value, the occurrence of an abnormal event for the block You can decide.

If the traffic state value for all blocks is less than a preset value, the device returns to the image input step, and determines whether to change lanes / traffic lines for blocks within the image, and calculates or updates traffic state values. The above-described process for determining the occurrence of the event can be continued.

On the other hand, when there are a plurality of mapped blocks in one image, the object detection unit 72, the traffic state calculation unit 73, and the event generation unit 74 for the corresponding block in the order of magnitude of the traffic state value of each block. The treatment of can be performed. Alternatively, when there are a plurality of the mapped blocks in one image, the corresponding blocks are assigned to the object detection unit 72, the traffic state calculation unit 73, and the event generation unit 74 in the order of magnitude of the traffic state value of each block. For blocks having the same traffic state value, a block close to a block having a larger traffic state value around it may be processed first.

In addition, the image analysis device 70 may further include an image acquisition unit (not shown) that acquires the road image. In addition, the image analysis device 70 may further include an operation unit that receives and stores the image or displays or receives information on the occurrence of the event. Such an operation unit may receive information on an object detected and recognized from the object detection unit 72 to generate and display traffic information, illegal parking information, speed violation information, or life crime prevention information.

When it is determined that the abnormal event has occurred, the image analysis device 70 may further include a communication unit for transmitting the abnormal event to an external system.

Meanwhile, the device illustrated in FIG. 7 may perform all the contents of the present invention described with reference to FIG. 6 or may be related thereto.

In addition, in the foregoing description, two values are compared as in S622. When comparing two values, an equal sign may be included in the magnitude relationship. That is, in the specification, "more" instead of "more" and "less than" instead of "less than" may be applied, and vice versa, which does not limit the scope of the present invention.

In the above specification, the "device" and its components (block mapping unit 71, object detection unit 72, traffic state calculation unit 73, event generation unit 74, etc.) perform the corresponding method or procedure. Although it has been described that the "device" and its components are only names, the scope of rights is not subordinate to it. That is, the method or procedure can be performed as a system as well as a device, and the method or method is not only implemented by software for detecting abnormal situations in roads and passageways, or by code that can be read by a computer or other machine or device. Can be done.

In addition, as another aspect of the present invention, the above-described proposal or operation of the invention is by means of a "computer" (a comprehensive concept including a system on chip (SoC) or a (micro) processor). Code that can be embodied, executed, or executed, or a computer-readable storage medium or computer program product that stores or includes the code, may be provided, and the scope of the present invention is to store the code or the code Including computer-readable storage media or computer program products.

Detailed description of the preferred embodiments of the present invention disclosed as described above has been provided to enable those skilled in the art to implement and implement the present invention. Although the above has been described with reference to the preferred embodiments of the present invention, those skilled in the art will understand that various modifications and changes can be made to the present invention described in the following claims. Thus, the invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (17)

delete An image analysis device for recognizing an abnormal situation, the image analysis device comprising:
A block mapping unit for mapping a road detected in the road image into blocks;
An object detection unit detecting a block in which an object enters or exits from among the blocks;
A traffic state calculator configured to calculate a traffic state value of a block adjacent to the detected block according to the entry or exit of the object in the detected block; And
Comprising an event generator configured to compare the traffic state value with a preset value, and when the comparison result satisfies a predetermined condition, determine that an event has occurred for a block that satisfies the predetermined condition,
The traffic state value includes a traffic state value of the neighboring block and a weight determined according to an entry or exit of an object in the detected block. The image analysis apparatus of claim 2, wherein when the object is detected as having advanced from the detected block, a weight applied to a block preceding the detected block is determined. The image analysis apparatus of claim 2, wherein when the object is detected as entering the detected block, a weight applied to a block after the detected block is determined. The image analysis apparatus of claim 2, wherein when the number of objects advancing from the detected block is greater than or equal to a threshold, a weight applied to a block preceding the detected block is determined. The image analysis apparatus of claim 2, wherein when the number of objects entering the detected block is greater than or equal to a threshold, a weight applied to a block after the detected block is determined. The image analysis apparatus of claim 2, wherein the traffic state value is based on at least one of an average moving speed of objects passing through a block or an average stopping time of objects in a block, and is obtained through a radar sensor or image processing. The method of claim 2,
When there are a plurality of the mapped blocks,
The image analysis apparatus, wherein the object detection unit, the traffic state calculation unit, and the event generation unit are processed for each block in order of magnitude of the traffic state value of each block. The method of claim 8,
For blocks having the same traffic state value, a block close to a block having a larger traffic state value around the block is first processed. The image analysis apparatus of claim 2, further comprising an image acquisition unit that acquires the road image. The method of claim 2, wherein the object detection unit:
An image analysis apparatus further comprising a radar sensor for detecting and tracking the object. The image analysis apparatus according to claim 2, further comprising an operation unit receiving and displaying the image, or receiving and displaying information on the occurrence of the event. The image analysis apparatus of claim 2, wherein the communication state value includes one of a plurality of values according to a section to which an average moving speed of an object in each block belongs. The image analysis apparatus of claim 2, wherein the traffic state value includes one of a plurality of values according to a section to which an average stop time of an object in each block belongs. The image analysis apparatus of claim 2, further comprising an operation unit receiving information on an object from the object detection unit and generating and displaying at least one of traffic information, illegal parking information, speed violation information, and life crime prevention information. As a method for recognizing an abnormal situation, the method is performed by an image analysis device,
Mapping the road detected in the road image to blocks;
Detecting a block in which an object enters or exits from among the blocks;
Calculating a traffic state value of a neighboring block of the detected block according to the entry or exit of the object in the detected block; And
Comparing the traffic state value with a preset value, and if the comparison result satisfies a predetermined condition, determining that an event has occurred for a block that satisfies the predetermined condition,
The traffic state value includes a traffic state value of the neighboring block and a weight determined according to an entry or exit of an object in the detected block. The method according to claim 16, wherein if the traffic state value for all blocks is less than a preset value, it returns to a road image input step.

Images