Obstacle detection support system

Abstract

The objective is to localize the range in which driver assistance is interrupted in an obstacle detection support system installed in a rail transport system. [Solution] The obstacle detection support system of the present invention comprises a rail detection unit and a branch position determination unit, wherein the branch position determination unit acquires rail information based on the rail detection results detected by the rail detection unit and identifies the start and end positions of the track branch. Furthermore, the branch position determination unit compares the track width of the lower edge (neighborhood information) and the track width of the upper edge (far information) in a processing unit and determines that there is a track branch if the track width of the upper edge is greater than or equal to the track width of the lower edge. Alternatively, the presence or absence of a track branch is determined by comparing the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far information) in a processing unit.

Description

[Technical Field]

[0001] This invention relates to an obstacle detection support system installed in a rail transport system that travels along a predetermined track. [Background technology]

[0002] In rail transport systems, where transport vehicles run on tracks, braking is required if there is an obstacle on the track. Therefore, detecting obstacles on the track is important for improving the safety and operability of rail transport systems. In recent years, research has been conducted on obstacle detection support systems that use external sensors such as millimeter-wave radar, laser radar, and cameras to detect obstacles on the track and support the driver's operation by notifying the driver of the detection results. Obstacle detection support systems suspend support near track junctions where false detections of obstacles are frequent. However, to reduce the burden on the driver, it is important to localize the area where support is not provided. Patent Document 1 discloses a road shape estimation device that estimates the shape of a road, including the determination of whether or not there is a junction. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-128612 [Overview of the project] [Problems that the invention aims to solve]

[0004] In obstacle detection support systems, driving assistance is suspended at railway junctions where false detection of obstacles is common. However, to reduce the burden on the driver, it is important to localize the areas where assistance is not provided. Patent document 1 describes a method for accurately estimating the shape of the road ahead of the vehicle's lane, and by determining whether the road widens as the vehicle moves in the direction of travel, the presence or absence of a junction is detected. Therefore, while the presence and starting position of a junction can be detected, the ending position of the junction cannot be detected, which has resulted in a delay in resuming driving assistance after it has been suspended. In order to address the above-mentioned problems, the present invention aims to localize the range in which driving assistance is interrupted in an obstacle detection support system installed in a rail transport system. [Means for solving the problem]

[0005] To solve the above problems, one representative obstacle detection support system of the present invention comprises a rail detection unit and a branch position determination unit, wherein the branch position determination unit acquires rail information based on the rail detection results detected by the rail detection unit and identifies the start and end positions of the track branch. [Effects of the Invention]

[0006] According to the present invention, in an obstacle detection support system installed in a rail transport system, the start and end positions of a track branch can be identified. By passing the identified results from the branch position determination unit to the obstacle detection unit, the driver assistance can be interrupted only from the start to the end position of the branch, enabling localization. Furthermore, by transmitting section information where the driver assistance display is interrupted from the obstacle detection unit to the driver assistance display unit, the driver can understand the range over which the driver assistance is interrupted. Other issues, configurations, and effects not mentioned above will be clarified by the following description of the embodiments. [Brief explanation of the drawing]

[0007] [Figure 1] This is a functional block diagram of the obstacle detection support system targeted by the first embodiment of the present invention. [Figure 2] This is a conceptual diagram illustrating the rail detection result of the rail detection unit in the first embodiment of the present invention. [Figure 3] This is a flowchart illustrating the process performed by the branch position determination unit in the first embodiment of the present invention. [Figure 4] This is a conceptual diagram illustrating the processing unit for determining branch position in the first embodiment of the present invention. [Figure 5] This is a conceptual diagram illustrating the acquisition of track width for branch position determination in the first embodiment of the present invention. [Figure 6] This is a conceptual diagram illustrating the case in the first embodiment of the present invention where the distance to the railway line is greater than or equal to that of the nearby line. [Figure 7] This is a conceptual diagram illustrating the acquisition of the number of rails for branch position determination in the first embodiment of the present invention. [Figure 8] This is a conceptual diagram illustrating the results of the branch location after branch location determination in the first embodiment of the present invention. [Modes for carrying out the invention]

[0008] [First Embodiment] Figure 1 is a functional block diagram of the obstacle detection support system targeted by the first embodiment of the present invention. This embodiment includes an external sensor 110 mounted on a train 100 that senses the external environment, and an obstacle detection support system 120 that notifies the driver of the presence or absence of obstacles ahead and supports the driver's operation. The obstacle detection support system 120 includes a rail detection unit 121 that detects rail information based on sensor information from the external sensor 110, a branch position determination unit 122 that determines the position of a branch based on the rail information, an obstacle detection unit 123 that detects obstacles in front of the train based on information from the external sensor 110, the rail detection unit 121, and the branch position determination unit 122, and a driving support display unit 124 that notifies the driver of support information based on the obstacle detection result information from the obstacle detection unit 123.

[0009] The external sensor 110 senses the state around (especially in front of) the train and transmits the sensing data to the rail detection unit 121 and the obstacle detection unit 123. The external sensor 110 includes a camera, LIDAR (laser range finder), millimeter-wave radar, GNSS (Global Navigation Satellite System), etc. The camera includes a monocular camera, a stereo camera, an infrared camera, etc. Generally, multiple sensors are installed for redundancy. When the train 100 travels in the forward direction, the external sensor installed on the leading vehicle on the forward direction side is used.

[0010] The rail detection unit 121 recognizes from the shape of the forward rail detected by image processing of the image obtained from the external sensor 110. Specifically, the method of detecting the rail by image processing includes a method of searching for the rail using the luminance gradient information in the image and a method of detecting the rail using semantic segmentation using DNN (Deep Neural Network). The rail detection result of the rail detection unit 121 will be described with reference to FIG. 2. In this embodiment, it is only necessary to be able to detect the rail, and the method is not limited. The rail detection unit 121 passes the obtained rail detection result to the branch position determination unit 122.

[0011] FIG. 2 is a conceptual diagram for explaining the rail detection result of the rail detection unit 121 in the first embodiment of the present invention.

[0012] The outer rectangular frame indicates the range of the image obtained from the external sensor 110. In a state where a track including a branch is displayed in the image, the thick line part represents the rail detection result detected by the rail detection unit 121 on the two-dimensional coordinates of the horizontal axis U and the vertical axis V.

[0013] The branch position determination unit 122 uses the rail detection result from the rail detection unit 121 to obtain the track width and the number of rails. Also, by using the rail information, the start position and the end position of the branch of the track are specified, and information is transmitted to the obstacle detection unit 123. Also, terms are defined as follows. Railway tracks: A pair consisting of two tracks. Rail: One rail

[0014] The obstacle detection unit 123 uses sensor information from the external sensor 110, the rail detection unit 121, and the branch position determination unit 122 to grasp the situation in front of the train and detect obstacles. The object detection processing of the obstacle detection unit 123 can utilize technologies used in the automotive field. For example, one method is to create a parallax image using a stereo camera and recognize the shape and position of objects ahead from the parallax image. Another method is to recognize objects on a monocular image using a DNN, or to recognize objects from LIDAR point cloud data. In this case, DNN is one of the methods used in machine learning, and it can detect various objects by extracting and learning the features of the target object.

[0015] In this embodiment, it is sufficient to detect obstacles, and the method of detection is not limited. The obstacle detection unit 123 makes a final determination of whether an obstacle exists and determines the section information to interrupt the driver assistance display based on information from the rail detection unit 121 and the branch position determination unit 122, as well as determination criteria from object detection results from a stereo camera, machine learning, and LIDAR.

[0016] The driver assistance display unit 124 informs the driver of the assistance provided to the driver via the HMI (Human Machine Interface) or via voice. In this embodiment, it is sufficient that the driver is informed of the assistance provided, including the interruption of the driver assistance display, and the method of such communication is not limited.

[0017] Next, the processing of the branch position determination unit 122 will be explained with reference to Figure 3. Figure 3 is a flowchart showing the process performed by the branch position determination unit 122 in the first embodiment of the present invention.

[0018] Step 301: The rail detection results received from the rail detection unit 121 are divided into processing units in order to determine the branch position. The division units will be explained later with reference to Figure 4.

[0019] Step 302: In step 301, branch position determination is performed for each processing unit divided into sections, and this branch position determination is repeated starting from the bottom of the image and ending at the top of the image or the vanishing point of the rail.

[0020] Step 303: In step 301, the track width is obtained for the target area (processing unit) that was divided. The process of obtaining the track width will be explained later, referring to Figure 5.

[0021] Step 304: In step 303, the track width of the lower edge (neighborhood information) and the track width of the upper edge (far-side information) of the processing unit obtained from the results obtained in step 303 are used to determine whether or not there is a branch in the target range (processing unit) divided in step 301. The track width of the lower edge and the upper edge are compared based on the criterion "Is the track width far-side ≥ neighborhood?". If the condition is met, the process proceeds to step 305; otherwise, the process proceeds to step 309. The case where the track width of the upper edge (far-side information) is greater than or equal to the track width of the lower edge (neighborhood information) will be explained later with reference to Figure 6. In this embodiment, the terms upper edge and far-side, and lower edge and neighborhood are used interchangeably.

[0022] Step 305: In step 304, if the track width on the upper edge (distant information) is greater than or equal to the track width on the lower edge (nearby information), the process of obtaining the number of rails is performed. The process of obtaining the number of rails will be explained later with reference to Figure 7.

[0023] Step 306: To determine whether or not there is a branch, the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) in the processing unit obtained in step 305 are used. For the determination criterion "number of rails ≥ 3", if the number of rails on the lower edge (neighborhood information) or the number of rails on the upper edge (far-away information) matches the condition, proceed to step 307; otherwise, proceed to step 308.

[0024] Step 307: To determine whether or not there is a branch, the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) obtained in step 305 are used. The criterion for determination is "Are the number of rails on the neighboring and far-away edges the same?" The number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) are compared. If the condition is met, the process proceeds to step 309; otherwise, the process proceeds to step 310.

[0025] Step 308: To determine whether or not there is a branch, the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) obtained in step 305 are used. The criterion for determination is "Are the number of rails on the neighboring and far-away edges the same?" The number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) are compared. If the condition is met, the process proceeds to step 310; otherwise, the process proceeds to step 311.

[0026] Step 309 If the determination in step 304 shows that the track width on the upper edge (distant information) is smaller than the track width on the lower edge (nearby information), or if the determination in step 306 shows that "number of rails ≥ 3" and the determination in step 307 shows that "the number of rails near and far are the same", then it is determined that there is no branch. Cases that meet these conditions will be explained later with reference to Figure 8.

[0027] Step 310 If the determination in step 306 is "Number of rails ≥ 3" and the determination in step 307 is "The number of rails near and far are different", or if the determination in step 306 is "Number of rails < 3" and the determination in step 308 is "The number of rails near and far are the same", then it is determined that there is a branch. Cases that meet these conditions will be explained later with reference to Figure 8.

[0028] Step 311 If the determination in step 306 is "Number of rails < 3" and the determination in step 308 is "Number of rails near and far are different", this condition is considered an abnormal case because it cannot occur due to the nature of the track.

[0029] The terms and processes that appeared in the description of each step above will be explained further below. First, the processing unit will be explained using Figure 4. Figure 4 is a conceptual diagram illustrating the processing unit for branch position determination in the first embodiment of the present invention.

[0030] The rectangle in the outer frame of the diagram represents the image area, and the thick lines represent the rail detection results received from the rail detection unit 121. The branch position determination unit 122 divides the image containing the rail detection results in the vertical direction based on a fixed number of section divisions. Here, the processing unit for the vicinity at the bottom of the image is denoted as K, and K increases as the image moves further away at the top of the image. Step 301 passes the divided content to step 302. An example of dividing the entire vertical image has been described, but it is also possible to divide only the area where rails are detected. By doing so, it is possible to narrow the vertical area after division, making it possible to identify the start and end positions of the track branch with greater accuracy. The number of divisions may also be changed according to the detection distance of the rails. For example, if rails are detected far away, the number of divisions may be increased. By doing so, it is possible to identify the start and end positions of the branch with greater accuracy. Examples of division numbers include 50, 100, and 200, but these are not the only examples.

[0031] Next, the acquisition of track width will be explained using Figure 5. Figure 5 is a conceptual diagram illustrating the acquisition of track width for branch position determination in the first embodiment of the present invention.

[0032] Based on the rail detection results for the target area for track width acquisition (see Processing Target (K) in Figure 5), the left coordinate of the bottom edge (U dl_k , V dl_k ), right coordinate (U dr_k , V dr_k ) and the left coordinate of the top edge (U ul_k , V ul_k ), right coordinate (U ur_k , V ur_kObtain . When obtaining the coordinates here, it shall be performed from both outer sides of the image of the target range. Next, obtain the track width (nearby information) 501 of the lower side and the track width (far - away information) 502 of the upper side from the left - and right - hand coordinates of the lower side and the upper side.

[0033] Next, the comparison of the track widths of the upper side and the lower side will be described using FIG. 6. FIG. 6 is a conceptual diagram for explaining the case where the far - away track width is greater than or equal to the nearby one.

[0034] When obtaining the coordinates in step 303, since it shall be performed from both outer sides of the image of the target range (refer to the processing target (K + 2) in FIG. 6), for the lower side, the left - hand coordinate is (U dl_k+2 , V dl_k+2 ), the right - hand coordinate is (U dr_k+2 , V dr_k+2 ), and for the upper side, the left - hand coordinate is (U )及び、上辺は左座標(U ul_k+2 , V ul_k+2 ), the right - hand coordinate is (U ur_k+2 , V ur_k+2 ur_k+2 ). Thus, the track width of the lower side (nearby information) 601 and the track width of the upper side (far - away information) 602 are obtained, and in the processing target (K + 2), the track width of the upper side (far - away information) is greater than or equal to the track width of the lower side (nearby information).

[0035] Next, the acquisition of the number of rails will be described using FIG. 7. FIG. 7 is a conceptual diagram for explaining the acquisition of the number of rails for branch - position determination in the first embodiment of the present invention.

[0036] Count how many rails there are on the lower side and the upper side of the target range (refer to the processing target (K + 2) in FIG. 7) including the rail for which the track width was obtained in step 303. In the case of the processing target (K + 2) in the figure, for the lower side, since it is only the rail targeted in step 303, the number is 2. On the other hand, for the upper side, in addition to the 2 rails targeted in step 303, there are 2 inner rails, so the total is 4.

[0037] Next, the determination of the presence or absence of a branch will be described using FIG. 8. FIG. 8 is a conceptual diagram for explaining the result of the branch location after branch - position determination in the first embodiment of the present invention.

[0038] If the determination in step 304 shows that the track width at the top (far information) is smaller than the track width at the bottom (nearby information), it represents the range of a normal track without branches, and corresponds to processing unit (K) in the diagram. Also, if the determination in step 306 is "number of rails ≥ 3" and the determination in step 307 is "number of rails near and far are the same", it represents a state where, for example, there is a branch before the image to be processed, and the track has already branched (a state without branches), and corresponds to processing units (K+3) and (K+4) in the diagram.

[0039] Furthermore, if the determination in step 306 is "Number of rails ≥ 3" and the determination in step 307 is "The number of rails near and far are different," it signifies the end of the track branching, which corresponds to the processing unit (K+2) in Figure 8. Also, if the determination in step 306 is "Number of rails < 3" and the determination in step 308 is "The number of rails near and far are the same," it signifies the start of the track branching or the middle of the branching, which corresponds to the processing unit (K+1) in the figure.

[0040] By repeating the above branch position determination, information on the presence or absence of a branch can be obtained for each processing unit divided in step 301, as shown in Figure 8, making it possible to identify the start and end positions of the track branch. In other words, according to this embodiment, by detecting not only the start of a branch but also the end of a branch (see processing unit (K+2)) and the subsequent state without a branch (see processing unit (K+3)), the range in which the driving assistance is interrupted can be localized, and it becomes possible to resume from the interruption more quickly.

[0041] In this embodiment, an example is described in which the start and end points of a track branch are identified using the results of branch position determination. The obstacle detection unit 123 detects the final obstacle, but by passing information on the start and end points of the branch from the branch position determination unit 122 to the obstacle detection unit 123, the driver assistance can be interrupted only from the start to the end point of the branch, or in abnormal cases, enabling localization. In addition, by transmitting the section information for which the driver assistance display is interrupted to the driver assistance display unit 124, the driver can understand the range in which the driver assistance display is interrupted.

[0042] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the present invention.

[0043] The following describes, but is not limited to, embodiments that may constitute the present invention. (Aspect 1) It is equipped with a rail detection unit and a branch position determination unit. The obstacle detection support system is characterized in that the branch position determination unit acquires rail information based on the rail detection result detected by the rail detection unit and identifies the start and end positions of the track branch. (Aspect 2) The obstacle detection support system according to Embodiment 1, characterized in that the branching position determination unit compares the track width of the lower edge (nearby information) and the track width of the upper edge (far information) in a processing unit, and determines that there is a track branch when the track width of the upper edge is equal to or greater than the track width of the lower edge. (Aspect 3) The obstacle detection support system according to embodiment 1 or 2, characterized in that the branching position determination unit identifies the presence or absence of a track branch by comparing the number of rails on the lower edge (nearby information) and the number of rails on the upper edge (far information) in the processing unit. (Aspect 4) The obstacle detection support system according to any one of embodiments 1 to 3, characterized in that the branch position determination unit detects not only the start position of a branch but also the end position of a branch by identifying whether or not there is a branch in the track for each processing unit. (Appendix 5) Equipped with an obstacle detection unit, The obstacle detection support system according to any one of embodiments 1 to 4, characterized in that the obstacle detection unit determines a section in which the operation support display for obstacle detection is interrupted based on the results of the start and end positions of the track branch from the branch position determination unit. (Aspect 6) Equipped with a driver assistance display unit, The obstacle detection support system according to embodiment 5, characterized in that the driving support display unit interrupts the driving support display to the driver based on section information that interrupts the driving support display from the obstacle detection unit. (Aspect 7) An obstacle detection support control method characterized in that, in the branch position determination unit, rail information is acquired based on the rail detection result detected by the rail detection unit to identify the start and end positions of the track branch. (Pattern 8) An obstacle detection support control method according to embodiment 7, characterized by including a step of comparing the track width of the lower edge (neighborhood information) and the track width of the upper edge (far-away information) in a processing unit, and determining that there is a track branch if the track width of the upper edge is greater than or equal to the track width of the lower edge. (Aspect 9) An obstacle detection support control method according to embodiment 7 or 8, characterized by including a step of identifying the presence or absence of track branching by comparing the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) in a processing unit. (Aspect 10) An obstacle detection support control method according to any one of embodiments 7 to 9, characterized in that it includes a step of detecting not only the start position of a branch but also the end position of a branch by identifying the presence or absence of a branch in the track for each processing unit. (Aspect 11) An obstacle detection support control method according to any one of embodiments 7 to 10, characterized in that a section is determined in which the obstacle detection driving support display is interrupted based on the results of the start and end positions of the track branch from the branch position determination unit. (Aspect 12) The obstacle detection support control method according to embodiment 11, characterized in that the driver assistance display to the driver is interrupted based on section information for which the driver assistance display is interrupted. [Explanation of Symbols]

[0044] 100... train 110... External sensor 120... Obstacle detection support system 121... Rail detection unit 122...Branch position determination unit 123... Obstacle detection unit 124...Driving support display unit

Claims

1. It is equipped with a rail detection unit and a branch position determination unit. The obstacle detection support system is characterized in that the branch position determination unit acquires rail information based on the rail detection result detected by the rail detection unit and identifies the start and end positions of the track branch.

2. The obstacle detection support system according to claim 1, characterized in that the branching position determination unit compares the lower track width (nearby information) and the upper track width (far information) in a processing unit, and determines that there is a track branch when the upper track width is equal to or greater than the lower track width.

3. The obstacle detection support system according to claim 1, characterized in that the branching position determination unit identifies whether or not there is a track branch by comparing the number of rails on the lower edge (neighborhood information) and the number of rails on the upper edge (far-away information) in the processing unit.

4. The obstacle detection support system according to claim 2 or 3, characterized in that the branch position determination unit detects not only the start position of a branch but also the end position of a branch by determining whether or not there is a branch in the track for each processing unit.

5. Equipped with an obstacle detection unit, The obstacle detection support system according to claim 1, characterized in that the obstacle detection unit determines a section in which the operation support display for obstacle detection is interrupted based on the results of the start and end positions of the track branch from the branch position determination unit.

6. Equipped with a driver assistance display unit, The obstacle detection support system according to claim 5, characterized in that the driving support display unit interrupts the driving support display to the driver based on section information that interrupts the driving support display from the obstacle detection unit.

Images