TWI453697B - The detection system of the driving space and its detection method - Google Patents

Description

Traveling space detection system and detection method thereof

The invention relates to a detection system for a travelable space and a detection method thereof, in particular to a detection system for determining a road travelable space by using stereo vision and a detection method thereof.

Traffic vehicles are closely related to people's lives, and cars are the primary personal transportation vehicles. However, car driving may result in possible traffic accidents due to different skills, experience and concentration. Therefore, today's cars are equipped with many electronic aids to assist and remind the car to avoid this situation.

Among them, the obstacle detection system is one of the directions that the automobile industry is striving for. The common obstacle detection system, such as the parking sensor, has become one of the standard equipments of many cars, effectively assisting the car to evaluate the parking space. Successfully completed the parking action. Recently, the special attention has been paid to the detection of obstacles in front of the car when it stops or travels. This kind of detection technology can avoid car accidents caused by ignoring obstacles in front of the car, and even in certain circumstances, it is a car. Driving speed reduction or braking greatly improves driving safety. However, how to effectively detect obstacles, the current related technologies still have many shortcomings to be improved.

Taking US Patent No. 5937079 "Method for Stereo Image Object Detection" as an example, the invention mainly determines the obstacle template of the input reference image by using the horizontal edge feature, and then obtains the position of the obstacle three-dimensional space through the histogram statistics. However, the case uses the horizontal features of the two-dimensional image as a model for obstacle comparison, which is susceptible to the design and quantity of the obstacle template.

In addition, in the U.S. Patent No. 6,801,244, "Obstacle Detection Apparatus and Method", the invention distinguishes the road surface from the obstacle by using the lane change matrix of the offline left and right camera images and using the parallax of the high-altitude image in the left and right images. . Although the case estimates the road travelable space by means of left and right image viewpoint conversion, this method is susceptible to road environment variation. For example, when a vehicle switches from a flat road to a slope road, it is misjudged to be an obstacle. Things.

Furthermore, in the US Patent No. US 2006/0095207 "Obstacle Detection Using Stereo Vision", the invention detects obstacles by using the edge and color information of the two-dimensional image, and then estimates the obstacle through stereo vision. The three-dimensional spatial information is used to estimate the road travelable space and the safe travel path. However, in this case, the obstacle is detected by the two-dimensional image feature, which is susceptible to the selection of the two-dimensional feature of the obstacle, that is, if the two-dimensional feature of the obstacle is not properly set in advance, the obstacle cannot be detected.

It can be understood from the above previous cases that in the past, in the road space detecting algorithm, most of the obstacles are detected first, and then the travelable space is estimated. Most of the obstacle detection algorithms are detected by image texture information (color/edge/shadow) or type information (length/width/aspect ratio) or template comparison, which is susceptible to environmental influence and low applicability. Obstacle obscuration and quantity also have an impact, which in turn creates errors in the path that can be traveled.

In addition, during the driving process, the response time required for the system to identify obstacles must be near-instant message feedback. The algorithm of the system can not only be interfered by other factors of the external environment, but also has a certain degree of reliability. The use of two sets of cameras to establish stereoscopic three-dimensional information can solve the influence of the external environment, but the huge computing load is the biggest bottleneck of this system, and also the key factor of whether the system is practical.

Accordingly, it is a primary object of the present invention to provide a detection system for determining a road travelable space using stereo vision.

Therefore, the detection system of the travel space of the present invention is mounted on a traffic vehicle and faces the moving direction of the traffic vehicle. The detection system comprises: a second image capturing unit, a processing unit and a memory unit. The image capturing units are disposed on the traffic carrier at intervals and are oriented in a direction in which the traffic vehicle advances to record a first image and a second image. The processing unit is electrically connected to the image capturing units. The memory unit is electrically connected to the processing unit, and stores the first image, the second image, and a detection program associated with the detected travelable space and executed by the processing unit, the image capturing unit and the processing unit And the memory unit cooperates to perform detection of the travelable space. The detection unit causes the processing unit to: first perform a stereoscopic image reconstruction operation, and convert the first image and the second image into a third image including a plurality of pixels, each pixel having a disparity value. Then, the third image is converted into an array of distances including a plurality of cells according to a road function. Furthermore, performing a cost function, estimating an obstacle value corresponding to the complex number by an obstacle item and a road plane term, wherein the obstacle item and the road plane item are based on the disparity values of the columns of the distance array And got it. Then, an optimized boundary estimation function is performed to calculate a smoothness value. Then, an optimization algorithm is executed according to the smoothness value to calculate a complex optimal travelable space boundary value.

Further, another object of the present invention is to provide a method for detecting a road travelable space by using stereoscopic vision.

Therefore, the method for detecting the driving space of the present invention operates on a detection system including a two-phase spaced image capturing unit, a memory unit and a processing unit, the detecting method comprising the following steps: First, the two images The capturing unit records a first image and a second image in the memory unit. Then, the processing unit performs a stereoscopic image reconstruction operation, and converts the first image and the second image into a third image including a plurality of pixels, each of the pixels having a disparity value. Moreover, the processing unit converts the third image into a distance array including a plurality of cells according to a road function. Then, the processing unit performs a cost function to estimate the obstacle value corresponding to the complex number by an obstacle item and a road plane item, wherein the obstacle item and the road plane item are based on the columns of the distance array. The difference is derived. Next, the processing unit performs an optimized boundary estimation function to calculate a smoothness value. Furthermore, an optimization algorithm is performed according to the smoothness value to calculate a complex optimal travelable space boundary value.

The effect of the invention is that the obstacle item and the road plane item in the cost function are obtained according to the disparity values of the columns of the distance array, and can be applied to different road situations, whether it is a plane road or an ups and downslope road. It has a good effect on the detection of obstacles.

The above and other technical contents, features and advantages of the present invention will be apparent from the following detailed description of the preferred embodiments.

Referring to Figures 1 and 2, the detection system of the travelable space of the present invention is mounted on a traffic vehicle 11. In the preferred embodiment, the traffic vehicle 11 is a vehicle, but is not limited thereto. The detection system includes a two-phase interval image capturing unit 21, a memory unit 22, a processing unit 23, a detecting unit 24, and a playing unit 25. The image capturing unit 21 is a camera in the preferred embodiment. Each image capturing unit 21 can capture a set angle of view image, such as a 30 degree view image, on the scene in front of the traffic vehicle 11 . The resolution of each image is set to 640 pixels × 480 pixels, that is, the column of the image has 640 pixels, and the column of the image has 480 pixels, but the resolution of the image is Not limited to this. The image capturing units 21 are installed at appropriate intervals of the traffic vehicle 11 at the same time as the headlights of the vehicle, such as at the front bumper, on the roof, above the central control panel, or through the windshield through the mechanism. The position is the same as the direction to be detected. In the preferred embodiment, the shooting direction is as shown by the arrow 12, moving toward the traffic carrier 11 in the forward direction, and the purpose is the traffic vehicle 11 When moving forward, detect if there are any obstacles in front of it and estimate the space that can be driven.

Referring to FIG. 1 and FIG. 3, the image capturing unit 21 on the left side can capture a first image 211, and the image capturing unit 21 on the right side can capture a second image 212, the first image 211 and the second image. Since the image 212 is slightly different in shooting angle, "parallax" is generated. The closer to the obstacle of the image capturing unit 21 in the first image 211 and the second image 212, the more obvious the parallax will be. On the contrary, the farther away from the image 撷Taking the obstacle of the unit 21, the case of parallax is less obvious. In the preferred embodiment, the obstacle in front of the moving direction of the traffic carrier 11 is a high wall 31 located on the left side of the traffic vehicle 11 , a locomotive 32 located on the left front side of the traffic vehicle 11 , and a traffic vehicle 11 . The bus 33 on the right front.

Referring to FIG. 1 and FIG. 2, the memory unit 22 stores the first image 211, the second image 212, and a detection program associated with detecting the travelable space and being executed by the processing unit 23, as shown in FIG. It is also possible to temporarily store the image files and calculation data required for the detection program for access by the detection program. In the preferred embodiment, the memory unit 22 is a memory module.

The processing unit 23 is electrically connected to the image capturing unit 21 and the memory unit 22. In the preferred embodiment, the processing unit 23 is a motherboard module including a central processing unit. The memory unit 22 and the processing unit 23 of the detection system are not limited to being implemented by a car computer (Car PC), and may be fabricated as a dedicated independent chip or a separate motherboard integrated into the electronic control system of the vehicle.

The detecting unit 24 is electrically connected to the processing unit 23, and the detecting unit 24 connects a directional light module of the traffic vehicle 11 and a vehicle speed module (not shown) in the preferred embodiment. The module can generate a steering information including a corresponding left turn signal and a right turn signal according to the opening condition of the left direction light and the right direction light, and the speed module generates a speed information according to the current speed. , such as 60 kilometers per hour. The steering information and the vehicle speed information are transmitted to the detecting unit 24, and the detecting unit 24 transmits a plurality of detection signals corresponding to the turning information and the vehicle speed information to the processing unit 23. It should be noted that the steering information including the left turn signal and the right turn signal is not limited to be provided by the above directional light module, or may be provided by a steering wheel. When the steering wheel is rotated counterclockwise and clockwise by a specific angle, Generate corresponding steering information.

The playing unit 25 is electrically connected to the processing unit 23. The playing unit 25 is disposed on the dashboard of the traffic carrier 11 in the preferred embodiment, and is a liquid crystal display screen including a speaker, providing the driver with driving Relevant graphical information and sound warnings.

The image capturing unit 21, the memory unit 22, the processing unit 23, the detecting unit 24, and the playing unit 25 cooperate to perform detection of the travelable space, and the detecting program causes the processing unit 23 to execute The steps are explained later.

Referring to Figures 2 and 3, the method for detecting the travelable space of the present invention operates in the above-described detection system, the detection system includes the image capture unit 21, the memory unit 22 including the detection program, and the execution of the The detecting unit 23, the detecting unit 24, and the playing unit 25, the detecting method comprises the following steps: refer to FIG. 2, 3, and 4. First, as shown in step 401, the image capturing unit 21 The first image 211 and the second image 212 are recorded in the memory unit 22. The first image 211 and the second image 212 can be used as a material for generating a stereoscopic image and analyzing the distance due to the parallax described above.

Then, as shown in step 402, the processing unit 23 performs a stereoscopic image reconstruction operation, and uses the first image 211 and the second image 212 to create a third image that exhibits an obstacle distance in the scene. In the preferred embodiment, the first image 211 and the second image 212 are converted into a third image including a plurality of pixels by using a feature point matching method, and each pixel has a view. Difference (Disparity). The feature point matching method refers to finding the same feature points (such as the locomotive 32) in the first image 211 and the second image 212, and confirming the view of each pixel in the third image according to the feature points. Difference. In the third image, the resolution is the same as the first image 211 and the second image 212, and the columns and columns of the image are 640 pixels and 480 pixels, respectively, but the third image is displayed in 16 gray scales. The disparity value of the pixel, the deeper the gray level of a pixel in the third image, the smaller the degree of parallax is, and the pixel is further away from the image capturing unit 21, and vice versa. The lighter the gray level of the pixel, the higher the degree of parallax, and the fact that the pixel is closer to the image capturing unit 21, and the block of each pixel set and the parallax value (grayscale) is close to the preferred embodiment. In the example, it may be a locomotive 32, a bus 33, and a sky 34. Therefore, the third image is substantially a three-dimensional coordinate image composed of three columns of information: a column of images, a column of images, and a disparity. In the field of computer graphics, the generation of the third image is not limited to the feature point matching method, and the third image may be obtained by another algorithm.

Referring to FIG. 2 and FIG. 4, further, as shown in step 403, the processing unit 23 calculates a road function according to all the pixels in the same column in the third image and the disparity values of the pixels. In the preferred embodiment, the method for calculating the road function is that the processing unit 23 sets a set of two-dimensional coordinates, wherein the horizontal axis coordinate is set as the parallax value of the third image, and the vertical axis coordinate is set as the third coordinate. In the column of the image, each pixel is filled into the vertical axis and the horizontal axis of the two-dimensional coordinate according to the "column" and "disparity value" to which each pixel belongs, to obtain an irregular curve formed by the pixels, and then Using a mathematical method of curve fitting to calculate the best curve closest to the above irregular curve to find the relationship between "column" and "disparity value" of each pixel, and the formula representing the above relationship is called The road function, and the road function is: column of the third image = (disparity value x road constant A) + road constant B. In the preferred embodiment, the road constant A is 0.6173, and the road constant B is 246.0254.

Then, as shown in step 404, the processing unit 23 converts the third image according to the road function described above, in order to obtain a conversion relationship between the column of the third image and the disparity value for subsequent operations. The vertical axis of the original third image is a column of the third image, and the horizontal axis is a column of the third image, and the processing unit 23 uses the road function to convert the column of the third image into a corresponding disparity value. The horizontal axis is maintained as a column, and the pixels of the original third image are rearranged according to the new coordinate system to generate a distance information.

Referring to Figures 2, 4, and 5, then, as shown in step 405, the processing unit 23 performs an Occupancy Grid conversion on the distance information to calculate a distance array 5 including a plurality of cells, substantially reducing the distance information. The amount of data increases the computational efficiency of the processing unit 23 to achieve an immediate processing effect. The so-called possessive transformation is to convert the original high-precision two-dimensional data into a set of lower-precision two-dimensional lattice arrays. In the preferred embodiment, the field of the distance information is 640 pixels (corresponding to the column of the original third image), and the vertical value of the vertical axis is 16 steps, for a total of 10240 groups of "parallax values". Assuming that the width of each of the distance arrays 5 is set to 40 pixels and the height is a 1st order disparity value, the horizontal axis of the array 5 is 16 cells, and the height is also 16 cells, for a total of 256 groups of columns. The parallax value, the greatly reduced amount of data helps to reduce the computational load of the processing unit 23. It must be noted that the purpose of the above-described possessive conversion is to reduce the amount of data to be processed, but is not limited to the above method.

The meaning of the distance array 5 can be regarded as a top view of the space in front of the traffic carrier 11 as shown in FIG. 1. Since the obstacle in front of the traffic vehicle 11 is closer to the traffic vehicle 11, the pixel corresponding to the obstacle is viewed. The larger the difference, the opposite, the farther away from the traffic vehicle 11, the smaller the pixel value corresponding to the obstacle, or even closer to zero. The grid labeled "X" in the distance array 5 represents a certain column, and the parallax values of the pixels are mostly gathered here, that is, there is an obstacle in the column at a specific distance.

2, 3, and 5, as shown in step 406, in order to further reduce the actual computing load of the processing unit, the processing unit 23 processes the corresponding image in the distance array 5 according to the detection signal of the detecting unit 24. The data of the complex detection areas 61, 62, 63 are shown. In the preferred embodiment, the detection signals are changed according to whether the speed information is higher than a preset speed, such as 30 kilometers, and the steering information. The detection signals are shown in Table 1:

In the case of the detection signal 1, the traffic vehicle 11 is traveling straight ahead and the vehicle speed is higher than the preset speed, and the obstacle affecting the travel must be in front of the traffic vehicle 11, so that only the detection is required. The portion shown in the area 62 is measured. In the case of the detection signal 2, the right direction light of the traffic vehicle 11 is turned on and the vehicle speed is higher than the preset speed, indicating that the traffic vehicle is about to be switched to the right lane, so it is necessary to detect as shown by the detection areas 62, 63. section. In the case of the detection signal 3, the left direction light of the traffic vehicle 11 is turned on and the vehicle speed is higher than the preset speed, which means that the traffic vehicle 11 is about to be switched to the left lane, so it is necessary to detect as shown in the detection areas 61, 62. section. In the case of the detection signal 4, mainly in a crowded section such as an urban area and the vehicle speed is lower than a preset speed, it is necessary to detect portions such as the detection areas 61, 62, 63 to comprehensively prevent possible Danger. In the preferred embodiment, it is assumed that the traffic carrier 11 is in the condition of the detection signal 4, so that all data in the distance array 5 must be processed.

It should be noted that the steering information indicated by the detection signals 2 and 3 is determined by the left and right turn signals corresponding to the left and right direction lights in the preferred embodiment, but is not limited thereto. It is also possible to detect whether the steering wheel of the traffic carrier 11 is rotated clockwise or counterclockwise by more than a specific angle, and accordingly generate corresponding left and right turn signals.

Referring to Figures 2, 4, and 5, further, as shown in step 407, the processing unit 23 performs a cost function to estimate a barrier value for each disparity value of each column in the array 5, The higher the obstacle value, the more likely it is that there is an obstacle. The cost function is:

v ( d , u )=ω ₁ × Object ( d , u )+ω ₂ × Road ( d , u )

Where v(d, u) is the obstacle value of the u-th column and the d-th disparity value in the distance array 5, ω ₁ is an obstacle item weight constant, and ω ₂ is a road plane item weight constant, The two weight constants are set to 30 and 50 respectively in the preferred embodiment to obtain a better detection effect, but are not limited thereto, and the two weight constants can be elastically adjusted according to the results of actual tests.

Object(d, u) is an obstacle item, that is, in the uth column of the distance array 5, the parallax values of the image capturing unit 21 to the obstacle change. Its function is:

Where v = v _min ~ v(d) means that v _{min is} the cost function to estimate the initial address of the obstacle item in the u column of the distance array 5, in the distance array 5, v _min refers to The highest disparity value, that is, the lowest column (column = 0) of the first image 211 or the second image 212 shown in FIG. ω represents a binary judgment function, where arg=d _u,v -d, if ω|arg| < a predetermined threshold, ω(arg)=1, if ω|arg|≧ the preset threshold, then ω(arg) = 0. In the preferred embodiment, the preset threshold is 20, but is not limited thereto.

Road(d, u) is a road plane term, that is to say, in the u column of the distance array 5, the parallax value of the obstacle to the farthest point changes. Its function is:

Where v = v(d) ~ v _max means that v _{max is} the cost function to estimate the end point address of the obstacle item in the u column. In the distance array 5, v _max refers to the lowest view. The difference is the uppermost column of the first image 211 or the second image 212 shown in FIG. 3 (column = column height of the image). ω is the same binary judgment function as described above, and is similarly judged based on the preset threshold value.

It is particularly noted that in the present invention, the road plane term uses the disparity value as a calculated parameter, since the front of the traffic vehicle 11 as shown in FIG. 1 is an uphill road or a downhill road, and the cost function is Words are not considered obstacles and increase the value of the obstacle.

Then, as shown in step 408, the processing unit 23 calculates an initial travelable space boundary value of each of the distance arrays 5 according to an initial boundary estimation function, and the initial travelable space boundary values. The function corresponding to I(u) is as follows.

Each column (u) of the distance array 5 initially makes the boundary value of the travelable space concatenated into a curve, and the initial boundary of the available travel space is not accurate enough, but the subsequent travelable space can be reduced. The search area of the boundary estimate relatively reduces the burden calculated by the processing unit 23 to increase the speed of obstacle recognition.

Next, as shown in step 409, the processing unit 23 calculates a smoothness value using the initial travelable space boundary value according to an optimized boundary estimation function. The purpose is to know whether there is an insufficiently smooth change between the initial travelable space boundary values of all the columns of the distance array 5, and if so, the initial travelable space boundary value representing a certain column of the distance array 5 may be due to noise. The influence of other factors causes it to differ greatly from the initial drivable space boundary value of another adjacent column. The optimized boundary estimation function is:

C _{u , d , k} = E ₁ ( u , d ) + E ₂ ( u , d , k )

Wherein, if the smoothness value C _{u,d,k is} higher, the problem that the difference between the adjacent columns is larger is more likely to exist, wherein E ₁ ( u,d )= v ( d,u ) represents the The energy value (Cost) of the u-th and d-th disparity values of the initial drivable space boundary, and E ₂ ( u,d,k )=ω ₃ ×(dk), representing the u-th column and the d-th parallax The difference between the initial drivable space boundary of the value and the energy value of the initial ejectable space boundary of the adjacent u+1 column and the kth disparity value, ω ₃ is a constant, which is set in the preferred embodiment as 0.5.

As shown in step 410, the processing unit 23 determines whether the smoothness value C _u,d,k is higher than a predetermined value to perform an optimization algorithm. If yes, the curve formed by the boundary values of the initial drivable spaces is not smooth, that is, the boundary value of the initial drivable space corresponding to at least one of the columns in the array 5 may be disturbed by noise, etc., and the adjacent column is generated. The initial drivable space boundary value is different. Therefore, as shown in step 411, the processing unit 23 calculates one of the optimal drivable space boundary values for each initial drivable space boundary value according to the optimization algorithm. . The optimization algorithm is a dynamic programming method in the preferred embodiment, but is not limited thereto, and other mathematical calculation methods can be used. Then, as shown in step 412, each of the optimal travelable space boundary values in the range array 5 is stored in the memory unit 22. The optimal travelable space boundary value corresponding to each column in the distance array 5 is a barrier line segment 52 as shown in FIG. 7, and the obstacle line segment 52 represents an obstacle and a traffic vehicle 11 as shown in FIG. the distance.

Returning to step 410, if the processing unit 23 determines that the smoothness value C _{u,d,k is} not higher than the preset value, the processing unit 23 does not need to perform the optimization algorithm described above, but as steps. At 412, each of the optimal travelable space boundary values in the range array 5 is stored in the memory unit 22.

Up to step 412, the information collected by the processing unit 23 about the distance array 5 is sufficient to know the distance between the traffic vehicle 11 and the obstacle as shown in FIG. 1. If the distance of the obstacle is too close, the playing unit can be utilized. The 25's horn sounded a warning to alert the driver. However, although the distance array 5 described above can provide the distance information between the traffic carrier 11 and the obstacle as shown in FIG. 1, the information presented in this way cannot be directly understood by the driver, and therefore must be changed. The coordinate system of the array 5, that is, correspondingly returns to the presentation of the first image 211 and the second image 212 as shown in FIG.

Referring to FIGS. 2, 7, and 8, then, as shown in step 413 of FIG. 4, the processing unit converts the distance array 5 into an obstacle detection layer 7 including a plurality of detection cells 71, and coordinates the coordinates of the array 5 The system (the horizontal axis is the column and the vertical axis is the disparity value) is converted into the coordinate system of the obstacle detection layer 7 (the horizontal axis is the same column as the distance array, and the vertical axis is the column), and at the same time, the best The drivable space boundary value is also correspondingly converted into a plurality of obstacle cue line segments 72. The way in which the disparity value is converted into a column can be calculated by using the road function calculated in step 403, and finally the obstacle detecting layer 7 including the detecting cells 71 as shown in FIG. 8 is calculated. Specifically, in the obstacle detection layer 7, the boundary value of the optimal travelable space is bounded, and the column is lower than the optimal travelable space boundary value, that is, the obstacle prompt line segment 72 is a travelable space. (Detection cell 71 marked "○"), the column is higher than the optimal travelable space boundary value as an obstacle area (detection cell 71 labeled "X"), and in the preferred embodiment The processing unit 23 processes only the detection cells 71 corresponding to the actual road portions of the obstacle detection layer 7, and in most cases, the detection cells 71 of the corresponding road portions form a trapezoidal region.

Referring to FIGS. 2, 8, and 9, further, as shown in step 414 of FIG. 4, the processing unit uses one of the first image 211 and the second image 212 as shown in FIG. 2 as a base image 8, the obstacle. The object detection layer 7 and the optimal travelable space boundary are superimposed on the base image 8 and display a composite image 9 in the playback unit 25 to provide the size and distance of the driver's interpretation obstacle of the traffic vehicle 11. Wait for tips. For example, in the composite image 9, the plurality of walls 31, the locomotive 32, and the bus 33 have a plurality of light bars representing the optimal drivable space boundary, and corresponding to the light bar of the obstacle cue line segment 72 as shown in FIG. 91, and the semi-transparent mask presented by the obstacle detection layer 7 under the light strip 91 represents the travelable area, corresponding to the detection grid 71 marked as "○" as shown in FIG. 7; conversely, there is no translucent cover The portion of the cover represents an obstacle area that cannot be driven, corresponding to the detection grid 71 labeled "X" as shown in FIG.

Referring to FIG. 2 and FIG. 4, the above steps 401 to 414 are performed in the preferred embodiment. The processing unit 23 is executed once every preset time, such as 1 second, so as to be able to grasp the traffic carrier every 1 second. The condition of the obstacle in the traveling direction is not limited thereto, and may be designed to be elastically adjusted according to the current speed of the traffic vehicle 11 or adjusted to a preset time in accordance with the calculation speed of the processing unit 23 in accordance with the efficiency.

Referring to Figures 1 and 2, since the present invention not only can detect obstacles, but also can further know the travelable space, it has the following applications:

1. When the traffic vehicle 11 encounters temporary obstacles on the roadside (vehicles parked on the shoulders and construction fences), or is overtaken by other vehicles and cut into the driving lane, the position of the obstacle can be immediately judged. Provide the corresponding driving guidance and advise the driver to the proper driving direction.

Second, it is applied to the stop of the traffic vehicle 11, such as roadside parking or reversing into the warehouse. Since the processing unit 23 can compare the size of the dockable space and the pre-stored traffic carrier 11, it can be determined whether the traffic carrier 11 can be parked in the dockable space.

In summary, the road plane item of step 407 of the present invention is calculated from the disparity value, and different road situations, such as a flat road, an uphill road, and a downhill road, can be applied. In addition, in step 408, the initial drivable space boundary value of each column in the distance array 5 is calculated, and the efficiency of searching for the optimal travelable space boundary value is improved, and the existing technology for detecting the travelable space and the obstacle is overcome. Therefore, the object of the present invention can be achieved.

The above is only the preferred embodiment of the present invention, and the scope of the invention is not limited thereto, that is, the simple equivalent changes and modifications made by the scope of the invention and the description of the invention are All remain within the scope of the invention patent.

11. . . Traffic vehicle

12. . . arrow

twenty one. . . Image capture unit

211. . . First image

212. . . Second image

twenty two. . . Memory unit

twenty three. . . Processing unit

twenty four. . . Detection unit

25. . . Play unit

31. . . High wall

32. . . locomotive

33. . . Bus

34. . . sky

401~414. . . step

5. . . Distance array

51. . . grid

52. . . Barrier line

61~63. . . Detection area

7. . . Obstacle detection layer

71. . . Detection grid

72. . . Obstacle prompt line segment

8. . . Base image

9. . . Synthetic image

91. . . Light bar

1 is a top view of a scenario illustrating a second image capture unit in a detection system for a travelable space of the present invention;

2 is a system block diagram showing a preferred embodiment of the detection system for the travelable space of the present invention;

3 is a schematic diagram showing a first image and a second image captured by the image capturing unit in the preferred embodiment;

4 is a system flow diagram illustrating a preferred embodiment of a method for detecting a travelable space of the present invention;

Figure 5 is a schematic diagram of a two-dimensional data illustrating a distance array in the preferred embodiment;

Figure 6 is a top plan view illustrating the complex detection area in the preferred embodiment;

7 is a schematic diagram of a two-dimensional data, illustrating that the complex number in the preferred embodiment corresponds to a barrier line segment corresponding to a boundary value of a plurality of optimal travelable spaces;

Figure 8 is a schematic diagram showing an obstacle detection layer in the preferred embodiment; and

FIG. 9 is a schematic diagram showing an image of the obstacle detection layer and the plurality of obstacle warning lines stacked on a base image and displaying a composite image.

401. . . The image capturing unit records the first image and the second image

402. . . Processing unit performs stereo image reconstruction operation

403. . . The processing unit calculates the road function

404. . . The processing unit converts the third image into distance information

405. . . The processing unit performs a possessive conversion on the distance information

406. . . The processing unit processes the distance array according to the dynamic condition

407. . . Processing unit execution cost function

408. . . The processing unit calculates the initial drivable space boundary value according to the initial boundary estimation function

409. . . The processing unit calculates a smoothness value according to an optimized boundary estimation function

410. . . The processing unit determines whether the smoothness value is higher than a preset value

411. . . The processing unit calculates the optimal travelable space boundary value according to the optimization algorithm

412. . . The processing unit stores the optimal travelable space boundary value in the storage unit

413. . . Processing unit conversion distance array is obstacle detection layer

414. . . The processing unit outputs the base image, the obstacle detection layer, and the optimal travelable space boundary as synthetic images

Claims (10)

A detection system for a travel space is mounted on a traffic vehicle and oriented toward a direction of movement of the traffic vehicle. The detection system includes: two image capture units, the image capture units are spaced apart from each other And a processing unit electrically connecting the image capturing unit; and a memory unit electrically connected to the processing unit, and facing the traffic vehicle in a forward direction And storing the first image, the second image, and a detection program associated with detecting the travelable space and executing by the processing unit, wherein the image capturing unit, the processing unit, and the memory unit cooperate to perform driving Detecting the space; the detecting program causes the processing unit to perform a stereoscopic image reconstruction operation, converting the first image and the second image into a third image including a plurality of pixels, each pixel having a disparity value according to The road function converts the third image into an array of distances including a plurality of lattices, performs a cost function, and estimates an obstacle corresponding to the complex object by an obstacle item and a road plane term. The obstacle item and the road plane item are obtained according to the disparity values of the columns of the distance array, performing an optimized boundary estimation function to calculate a smoothness value, and performing according to the smoothness value An optimization algorithm is used to calculate the boundary value of the complex optimal travelable space. The detection system of the travelable space according to claim 1 further includes a detection unit electrically connected to the processing unit and outputting a plurality of detection signals, wherein the equal-number detection area of the distance array is The processing unit determines to process one of the detection regions of the distance array according to the detection signals. According to the detection system of the travelable space described in claim 1, wherein the processing unit converts the third image into the distance array by converting the ordinate axis of the third image from the column to the parallax Value, then perform a possessive conversion. The detection system of the travelable space according to claim 1, wherein the cost function is each of the obstacle values = an obstacle item weight constant × the obstacle item + a road plane item weight constant × the road plane item. According to the detection system of the travelable space described in claim 1, the processing unit performs an initial boundary estimation function before calculating the boundary estimation function, and calculates the complex corresponding distance by using the obstacle values. The initial drivable space boundary value of the column of the array to calculate the smoothness value. A method for detecting a travelable space, comprising: a two-phase spaced image capture unit, a memory unit and a processing unit detection system, the detection method comprising the following steps: (a) the second image capture The unit records a first image and a second image in the memory unit; (b) the processing unit performs a stereoscopic image reconstruction operation, and converts the first image and the second image into a third image including a plurality of pixels, each of The pixel has a disparity value; (c) the processing unit converts the third image into a distance array including a plurality of cells according to a road function; (d) the processing unit performs a cost function to an obstacle item and A road plane term estimates an obstacle value corresponding to the complex number, wherein the obstacle item and the road plane item are obtained according to a disparity value of each column of the distance array; (e) the processing unit performs an optimization a boundary estimation function to calculate a smoothness value; and (f) performing an optimization algorithm based on the smoothness value to calculate a complex optimal travelable space boundary value. According to the method for detecting the travelable space described in claim 6 of the patent application, between the steps (c) and (d), a step (g) is further included, and the processing unit detects the signal according to the complex number of the detecting unit. The processing unit determines one of the complex detection regions of the distance array to be processed according to the detection signals. According to the detection system of the travelable space described in claim 6, wherein in the step (c), the processing unit converts the ordinate axis of the third image from the column to the disparity value, and then executes one Ownership conversion. According to the detection system of the travelable space described in claim 6, wherein in the step (d), the cost function is each of the obstacle values = an obstacle item weight constant x the obstacle item + a road Plane item weight constant × the road plane item. The detection system of the travelable space according to claim 9 of the patent application, in the step (d) and the step (e), further comprising a step (h), the processing unit performing an initial boundary estimation function, The initial drivable space boundary values corresponding to the columns of the distance array are calculated using the obstacle values, and in the step (e), the processing unit calculates the smoothness values using the initial drivable space boundary values.