Method for estimating distance between shielded object in front of vehicle and collision warning system

Abstract

The invention relates to a method for estimating the distance of a blocked object in front of a vehicle, which is used for solving the problem that the known distance detection of the front vehicle is interfered by blocking to misjudge the distance. The method comprises the following steps: capturing an image in front of a vehicle; identifying a plurality of objects in the image and generating a plurality of corresponding object frames; selecting one of the plurality of object frames as a target object frame and a plurality of other reference object frames in turn; calculating a relative distance between an object corresponding to the target object frame and the vehicle according to the overlapping condition of each reference object frame and the target object frame; and selecting the next object frame as a new target object frame and calculating the relative distance. The invention further discloses a collision warning system. Therefore, the effects of avoiding misjudgment of distance and accurately preventing collision accidents can be achieved.

Description

Technical Field

[0001] This invention relates to a driver assistance technology, and more particularly to a method for accurately judging the distance to a vehicle ahead and improving driving safety by estimating the distance to an obstructed object in front of the vehicle, and a collision warning system for executing the method. Background Technology

[0002] Advanced Driver Assistance Systems (ADAS) use various sensors, such as radar, ultrasound, and cameras, to detect changes in the vehicle's surroundings in real time. They can locate the driving path, detect obstacles, and calculate the distance to the vehicle in front, thus warning and assisting the driver in dealing with various driving situations, thereby improving driving safety, reducing accidents, and increasing driving convenience.

[0003] The aforementioned known driver assistance technologies capture images of the area in front of the vehicle, analyze the types, actual sizes, and relative positions of objects in the image, and then estimate the actual distance to the objects based on camera parameters. This allows them to determine whether the vehicle is moving away from or close to the objects and whether a collision is likely to occur. However, when multiple objects overlap in the image, it can affect the object recognition results and cause errors in the estimated distance, leading to collision warning function malfunctions. This can cause unnecessary interference to the driver and even lead to traffic accidents.

[0004] In view of this, existing distance estimation and collision warning technologies still need to be improved. Summary of the Invention

[0005] To address the aforementioned problems, the purpose of this invention is to provide a method for estimating the distance to an obstructed object in front of a vehicle, which can accurately calculate the relative distance in front of the vehicle.

[0006] A secondary objective of this invention is to provide a collision warning system that can accurately predict the time of a collision.

[0007] Another object of the present invention is to provide a collision warning system that can avoid false warnings.

[0008] Another object of the present invention is to provide a collision warning system that can improve driving convenience.

[0009] The directions or similar terms used throughout this invention, such as "front," "left," "right," "top," "bottom," and "side," are mainly for reference to the directions in the accompanying drawings. These directions or similar terms are only used to assist in explaining and understanding the various embodiments of this invention and are not intended to limit this invention.

[0010] The use of the quantifiers “a” or “an” in this specification is for convenience only and to provide the general meaning of the scope of the invention; in the context of the invention, it should be interpreted as including one or at least one, and the concept of a single also includes the plural, unless it clearly means otherwise.

[0011] The present invention provides a method for estimating the distance of an obstructed object in front of a vehicle, comprising: capturing an image of the front of a vehicle; identifying several objects in the image and generating several corresponding object frames; sequentially selecting one of the object frames as a target object frame; comparing the bottom edge position of the selected target object frame with the bottom edge position of the other object frames, and defining each object frame whose bottom edge position is lower than that of the target object frame as a reference object frame; determining whether each reference object frame overlaps with the target object frame; selecting a distance calculation method based on the overlap between each reference object frame and the target object frame, calculating a relative distance between the object corresponding to the target object frame and the vehicle; and selecting the next object frame as the new target object frame and calculating the relative distance, until the relative distance of each object frame is calculated.

[0012] The collision warning system of the present invention includes: a camera module located on a vehicle, capturing an image of the front of the vehicle; a computing module connected to the camera module and receiving the image, the computing module performing the distance estimation method described above, calculating a relative distance of an object in the image, recording several consecutive relative distances at a time interval, calculating the time change rate of the relative distance to predict a collision time; and an alarm module connected to the computing module and receiving the collision time, issuing a collision warning when the collision time is less than a warning threshold and it is determined that the object is in the vehicle's driving path.

[0013] Accordingly, the distance estimation method for obstructed objects in front of a vehicle and the collision warning system for performing the estimation method of the present invention can accurately calculate the correct distance between the vehicle and each of the objects by analyzing the mutual obstruction of several objects in the image in front of the vehicle. Then, the collision warning system calculates and judges whether each of the objects will collide with the vehicle and accurately predicts the time of collision. It has the effects of avoiding false warnings, preventing collision accidents and improving driving convenience.

[0014] Specifically, if the bottom edge of the selected target object frame is lower than the bottom edges of several other object frames, then no reference object frame is set, and the relative distance is calculated based on the height or width of the target object frame. This allows for the determination of the object corresponding to the target object frame that is closest to the vehicle, enabling accurate distance estimation without obstruction and improving computational efficiency.

[0015] The overlap between the reference object frames and the target object frame includes the following: the left or right side of each reference object frame overlaps with the target object frame, indicating that the width direction of the target object frame is obscured; the top edge of each reference object frame overlaps with the target object frame, indicating that the height direction of the target object frame is obscured. This allows for confirmation of the occlusion status of the target object frame, avoiding the use of incorrect reference values ​​and improving the accuracy of distance estimation.

[0016] The distance calculation method includes: when the width direction of the target object frame is occluded, the relative distance is calculated based on the height of the target object frame; when the height direction of the target object frame is occluded, the relative distance is calculated based on the width of the target object frame; when the target object frame is not occluded, the relative distance is calculated based on either the height or width of the target object frame; when both the width and height directions of the target object frame are occluded, the calculation of the relative distance is cancelled. Thus, calculations are performed only for unoccluded objects or parts of the frame, avoiding the calculation of relative distances for objects with insufficient information or no immediate collision risk, thereby reducing invalid calculations and lowering the computational burden.

[0017] The formula for calculating the relative distance is as follows: Where D is the relative distance in meters; F is the pixel focal length of the camera device in pixels; H is the actual height or width of the object corresponding to the target object frame in meters; and M is the height or width of the target object frame in pixels. This allows for object identification and calculations based on actual dimensions, improving calculation accuracy.

[0018] Wherein, the rate of change of the relative distance over time is a relative velocity, and the formula for the relative velocity is as follows: Where v is the relative velocity; i and j are natural numbers, i ≠ j; D i D represents the relative distance described in the i-th entry; j Let be the relative distance of the j-th record; and t be the time interval. Thus, recording several consecutive relative distances allows for the calculation of their changes over time, effectively determining the vehicle's movement.

[0019] Wherein, if the relative velocity is less than zero, the object approaches the other object. The collision time of the object is calculated using the following formula: Where B is the collision time; D is the relative distance; and v is the relative velocity. In this way, the collision time can be predicted, thus preventing collisions.

[0020] Wherein, the rate of change of the relative distance over time is a relative acceleration, and the formula for the relative acceleration is as follows: Where n, i, and j are natural numbers, i ≠ j, and n ≧ 2; D n v represents the relative distance mentioned in the nth entry; n The relative velocity is recorded at the nth relative distance; t is the interval time; and a is the relative acceleration. Thus, the more relative distances recorded, the more the relative acceleration can be calculated, enabling accurate determination of vehicle movement trends.

[0021] The collision time of the object is calculated using the following formula: Where B is a positive real number, then B represents the collision time; when B is a complex number or less than zero, no collision will occur. Thus, the collision time can be predicted, which has the effect of preventing collisions. Attached Figure Description

[0022] Figure 1 A schematic diagram of a collision warning system according to a preferred embodiment of the present invention is shown.

[0023] Figure 2 This diagram illustrates an object identification scenario using a distance estimation method according to a preferred embodiment of the present invention.

[0024] Figure 3 This diagram illustrates the target selection scenario for performing the distance estimation method according to a preferred embodiment of the present invention.

[0025] Figure 4 This diagram illustrates the distance calculation scenario when performing a distance estimation method according to a preferred embodiment of the present invention.

[0026] The components include: 1. Camera module, 2. Calculation module, 3. Alarm module, R. Image, B. Collision time, C. Object bounding box, C. T .Target object bounding box, C R . Reference object frame, U. bottom edge, h. height, w. width. Detailed Implementation

[0027] To make the above and other objects, features and advantages of the present invention more apparent and understandable, preferred embodiments of the present invention are described below in detail with reference to the accompanying drawings; in addition, those symbols that are used in different drawings are considered to be the same and their descriptions will be omitted.

[0028] Please refer to Figure 1As shown, this is a preferred embodiment of the collision warning system of the present invention, which includes a camera module 1, a calculation module 2 and an alarm module 3, wherein the calculation module 2 is connected to the camera module 1 and the alarm module 3 respectively.

[0029] The camera module 1 is preferably located in front of a vehicle and is positioned so that it is shooting in the direction the vehicle is traveling, capturing the forward field of view of the vehicle as an image R.

[0030] The calculation module 2 receives the image R to calculate a relative distance between an object in the image R and the vehicle. The calculation module 2 can calculate the relative distance once at each time interval to record the changes in the relative distance over several consecutive periods. The calculation module 2 can also calculate the rate of change of the relative distance over time and use it to predict a collision time B between the object and the vehicle.

[0031] Please refer to again Figure 2-4 The diagram illustrates the execution of the distance estimation method for an obscured object in front of a vehicle by the computation module 2. The steps include: capturing the image R in front of the vehicle; generating several object boxes C in the image R; and selecting a target object box C from the several object boxes C. T And several other reference object frames C R ; Determine each of the aforementioned reference object boxes C R With the target object frame C T Based on the overlap condition, a distance calculation method is selected to calculate a relative distance; and the next object box C is selected as the new target object box C. T And calculate the relative distance.

[0032] Please refer to Figure 2 As shown, image recognition can be performed on the image R to detect several objects, including cars, motorcycles, pedestrians, etc., moving in front of the vehicle, and each of the objects is marked with an object box C.

[0033] Please refer to again Figure 3 As shown, the bottom edge U of each of the object boxes C is marked, and one of the object boxes C is selected as the target object box C. T Compare the selected target object bounding box C T The bottom edge U position of each of the other object frames C is lower than that of the target object frame C. T If the bottom edge U is at a certain position, then the object frame C is set as a reference object frame C. R .

[0034] Please refer to Figure 3 and4 As shown, the next step is to determine each of the reference object boxes C. R Does it overlap with the target object frame C? T When each of the aforementioned reference object boxes C R The left or right side overlaps with the target object frame C T Then determine the target object bounding box C. T The width direction is obscured; when each of the reference object frames C R The top edge overlaps with the target object frame C. T Then determine the target object bounding box C. T The height direction is obstructed, such as Figure 3 As shown, the target object bounding box C T The width direction is defined by the reference object frame C. R To obstruct or block. Also, as in... Figure 4 As shown, the selected target object bounding box C T Neither the width nor the height of the reference object is affected by any of the reference object frames C. R Occlusion allows for the complete acquisition of the target object bounding box C. T The height h and width w.

[0035] Based on each of the aforementioned reference object frames C R Occlude the target object frame C T The overlap condition is considered when selecting the distance calculation method, including: when the target object bounding box C... T When the width direction is occluded, the target object frame C T The relative distance is calculated based on the height h; when the target object frame C T When the height direction is obstructed, the target object frame C T The relative distance is calculated based on the width w; when the target object bounding box C T When not obscured, the target object frame C is used. T The relative distance is calculated based on the height h or width w of the target object frame; when the target object frame C T When both the width and height directions are obscured, the calculation of the relative distance is cancelled.

[0036] Furthermore, when the selected target object box C T If the bottom edge U position of the reference object is lower than the bottom edge U positions of all other object frames C, then the reference object frame C is not set. R The target object bounding box C can be determined. T Unobstructed, with the target object frame C T The relative distance is calculated using either the height h or the width w.

[0037] The formula for calculating the relative distance is as follows: Where D is the relative distance, and C is the target object bounding box. T The corresponding actual distance between the object and the vehicle is in meters; F is the pixel focal length of the camera device, in pixels; H is the bounding box of the target object C. T The corresponding actual height or width of the object, in meters; M is the bounding box of the target object C. T The height h or width w is in pixels.

[0038] The distance estimation method for an obstructed object in front of a vehicle according to the present invention, after calculating a relative distance, will continue to select the next object bounding box C as the new target object bounding box C. T The same steps are then followed to determine and calculate the corresponding target object bounding box C. T The relative distance is calculated, and the steps are repeated until the relative distance of each of the object boxes C is calculated.

[0039] Please refer to Figure 1 As shown, the calculation module 2, using the aforementioned distance estimation method, can continuously monitor the relative distance between each object and the vehicle in the image R. When the calculation module 2 continuously records several relative distances, it can further calculate a relative speed between the object and the vehicle. The formula for the relative speed is as follows: Where v is the relative velocity; i and j are natural numbers, i ≠ j; D i D represents the relative distance described in the i-th entry; j Let be the relative distance of the j-th measurement; t be the interval time, which is the time difference between the monitoring times of two adjacent relative distance measurements. The unit of the relative speed is preferably meters per second, corresponding to the unit of the interval time in seconds.

[0040] If the calculated relative velocity is greater than zero, the object moves away from the vehicle; if the calculated relative velocity is equal to zero, the object maintains the relative distance from the vehicle; if the calculated relative velocity is less than zero, the object moves closer to the vehicle. The collision time B between the object and the vehicle can be calculated using the following formula: Where D is the relative distance; v is the relative velocity.

[0041] The alarm module 3 receives the collision time B and compares it with a preset warning threshold. If the collision time B is less than the warning threshold, it can be determined that the object poses a risk of colliding with the vehicle. The warning threshold is preferably set based on the driver's reaction speed, and can be 1.2 seconds to 3 seconds. However, this invention is not limited to this. Furthermore, as... Figure 2-4 As shown, the alarm module 3 can also determine whether an object is on the vehicle's driving path by the relative position of each of the corresponding object frames C and the vehicle's driving path. The vehicle's driving path can be limited to the distance from the center of the optical axis of the camera device to the width of the vehicle on both sides, extending forward. This extended distance can be 80 to 100 meters, but the invention is not limited to this. When the collision time B is less than the warning threshold, and it is determined that the object is on the vehicle's driving path, the alarm module 3 can issue a collision warning to remind the driver of an impending collision, or activate the driver assistance function to intervene and prevent the collision.

[0042] In another preferred embodiment, when the calculation module 2 records more than two relative distances consecutively, in addition to calculating the relative velocity of the object, it can further calculate the relative acceleration between the object and the vehicle when the cumulative relative distance reaches more than 15. Furthermore, the present invention is not limited to the aforementioned cumulative number of 2 or 15. The formula for the relative acceleration is as follows: Where n, i, and j are natural numbers, i ≠ j, and n ≧ 2; D n v represents the relative distance mentioned in the nth entry; n The relative velocity is recorded at the nth relative distance; t is the interval time; a is the relative acceleration.

[0043] The collision time B between the object and the vehicle can be calculated based on the relative acceleration system, using the following formula: Wherein, when B is a positive real number, B represents the collision time; when B is a complex number or less than zero, no collision will occur. Calculating the collision time B using the relative acceleration avoids situations where large variations in relative velocity cause the relative velocity at a certain point in time to deviate from the overall trend, leading to misjudgments of the calculated collision time. Recording a sufficient number of relative distances and calculating the relative acceleration and collision time B can increase the accuracy and reliability of the collision warning system.

[0044] In summary, the distance estimation method for obstructed objects in front of a vehicle and the collision warning system executing the method of the present invention can accurately calculate the correct distance between the vehicle and each of the objects by analyzing the mutual obstruction of several objects in the image in front of the vehicle. Then, the collision warning system calculates and judges whether each of the objects will collide with the vehicle and accurately predicts the time of collision. It has the effects of avoiding false warnings, preventing collision accidents and improving driving convenience.

[0045] Although the present invention has been disclosed using the above preferred embodiments, it is not intended to limit the present invention. Any modifications and alterations made by those skilled in the art to the above embodiments without departing from the spirit and scope of the present invention shall still fall within the technical scope protected by the present invention. Therefore, the scope of protection of the present invention shall include all changes within the meaning and equivalent scope of the appended claims.

Claims

1. A method for estimating the distance to an obstructed object in front of a vehicle, characterized in that, Include: Capture an image of the front of a vehicle; Several objects are identified in the image and corresponding object bounding boxes are generated; Select one of the several object bounding boxes in turn as a target object bounding box; Compare the bottom edge positions of the selected target object frame with those of several other object frames, and define each object frame whose bottom edge position is lower than that of the target object frame as a reference object frame; Determine whether each of the reference object frames overlaps with the target object frame; Based on the overlap between each of the reference object frames and the target object frame, a distance calculation method is selected to calculate a relative distance between the object corresponding to the target object frame and the vehicle. and Select the next object box as the new target object box and calculate the relative distance, until the relative distance of each object box is calculated.

2. The method for estimating the distance to an obstructed object in front of a vehicle as described in claim 1, characterized in that, If the bottom edge of the selected target object frame is lower than the bottom edges of several other object frames, then no reference object frame is set, and the relative distance is calculated based on the height or width of the target object frame.

3. The method for estimating the distance to an obstructed object in front of a vehicle as described in claim 1, characterized in that, The overlap between each of the reference object frames and the target object frame includes: the left or right side of each of the reference object frames overlaps with the target object frame, thus obscuring the width direction of the target object frame; the top edge of each of the reference object frames overlaps with the target object frame, thus obscuring the height direction of the target object frame.

4. The method for estimating the distance to an obstructed object in front of a vehicle as described in claim 3, characterized in that, The distance calculation method includes: when the width direction of the target object frame is obscured, the relative distance is calculated based on the height of the target object frame; when the height direction of the target object frame is obscured, the relative distance is calculated based on the width of the target object frame; when the target object frame is not obscured, the relative distance is calculated based on either the height or the width of the target object frame; when both the width and height directions of the target object frame are obscured, the calculation of the relative distance is cancelled.

5. The method for estimating the distance to an obstructed object in front of a vehicle as described in claim 2 or 4, characterized in that, The formula for calculating the relative distance is as follows: ; Wherein, D is the relative distance in meters; F is the pixel focal length of the camera device in pixels; H is the actual height or width of the object corresponding to the target object frame in meters; and M is the height or width of the target object frame in pixels.

6. A collision warning system, characterized in that, Include: A camera module, located on a vehicle, captures an image of the front of the vehicle; A computing module is connected to the camera module and receives the image. The computing module executes the distance estimation method for an obstructed object in front of a vehicle as described in any one of claims 1 to 5, calculates a relative distance of an object in the image, records a number of consecutive relative distances at an interval, calculates the time change rate of the relative distance, and uses it to predict a collision time. and An alarm module is connected to the computing module and receives the collision time. When the collision time is less than a warning threshold and it is determined that the object is in the vehicle's driving path, a collision warning is issued.

7. The collision warning system as described in claim 6, characterized in that, The rate of change of the relative distance over time is a relative velocity, and the formula for the relative velocity is as follows: ; Where v is the relative velocity; i and j are natural numbers, i ≠ j; D i D represents the relative distance described in the i-th entry; j The relative distance is the j-th reference; t is the interval time.

8. The collision warning system as described in claim 7, characterized in that, If the relative velocity is less than zero, the objects move closer together. The collision time of the objects is calculated using the following formula: ; Where B is the collision time; D is the relative distance; and v is the relative velocity.

9. The collision warning system as described in claim 6, characterized in that, The rate of change of the relative distance over time is a relative acceleration, and the formula for the relative acceleration is as follows: ; Where n, i, and j are natural numbers, i ≠ j, and n ≧ 2; D n v represents the relative distance mentioned in the nth entry; n The relative velocity is recorded at the nth relative distance; t is the interval time; a is the relative acceleration.

10. The collision warning system as described in claim 9, characterized in that, The collision time of the object is calculated using the following formula: ; Wherein, when B is a positive real number, B is the collision time; when B is a complex number or less than zero, no collision will occur.

Images