CN116704446B - Real-time detection method and system for foreign objects on airport runway pavement - Google Patents

Abstract

The invention discloses a real-time detection method of foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching, which includes the following steps: S1. Input an infrared image of the road surface scanned by an infrared laser line, and extract the laser line in the infrared image of the road surface; S2. Perform multi-neighbor multi-step gradient calculation on the extracted y-coordinate of the laser line to obtain the multi-step gradient curve corresponding to the points on the laser line; S3. Find the minimum and maximum values on the multi-step gradient curve, according to The characteristics of the image zero-crossing points are defined as the following characteristic patterns of foreign objects, pits and roadbed. S4. Filter misidentified foreign objects by calculating the height information of the foreign objects, and obtain the position of the foreign object according to the x-coordinate of the marked foreign object; S5. Repeat the operations from S1 to S4 for each frame of subsequent input image, and convert the The foreign object images are spliced to synthesize the final three-dimensional pavement information and foreign object marking results. The invention can detect foreign objects on airport roads efficiently, accurately and in real time.

Description

Real-time detection method and system for foreign objects on airport runway pavement

Technical field

The invention relates to the field of image processing for target detection, and in particular to a real-time detection method and system for foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching.

Background technique

Traditional foreign object detection methods usually rely on visual inspections and manual cleaning to check foreign objects on runways and airport surfaces. However, these methods cannot accurately identify foreign objects in a timely manner, and they also have great limitations in terms of maintenance time and cost. For this reason, real-time road foreign object detection technology came into being. The technology can monitor airport surfaces in real time as aircraft take off and land, and promptly identify and report potentially hazardous items. With the continuous development of drones, automated robots and AI technology, real-time FOD monitoring technology has become a key performance indicator for airport and aircraft maintenance, and is the only way to significantly improve aircraft, safety performance and efficiency. Currently, there are some technical solutions involving real-time foreign object detection on the market, but these solutions are not perfect enough, and their detection accuracy and performance still need to be further improved.

Contents of the invention

The main purpose of the present invention is to provide a real-time detection method and system for foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching that can improve detection accuracy.

The technical solution adopted by the present invention is:

A real-time detection method of foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching is provided, including the following steps:

S1. Input the infrared image of the road surface scanned by the infrared laser line, and extract the laser lines in the infrared image of the road surface;

S2. Perform multi-neighbor multi-step gradient calculation on the extracted y-coordinate of the laser line to obtain the multi-step gradient curve corresponding to the points on the laser line;

S3. Find the minimum and maximum values on the multi-step gradient curve, and define the following characteristic patterns of foreign objects, pits and roadbed according to the characteristics of the zero-crossing point of the image:

On a multi-step gradient curve, if the gradient minimum appears first and then the gradient maximum appears within a certain window, this characteristic pattern represents foreign matter. The laser line between the gradient minimum and the gradient maximum The category label of the point on is set to Foreign Object;

If the gradient maximum value appears first, and then the gradient minimum value appears within a certain window range, then this feature mode is a pit, and the category label of the point on the laser line between the gradient maximum value and the gradient minimum value is set. for pits;

The remaining feature patterns are roadbed, and the category labels of the points on these laser lines are set to roadbed;

S4. Filter misidentified foreign objects by calculating the height information of the foreign objects, and obtain the position of the foreign objects based on the x-coordinate of the marked foreign objects;

S5: Repeat the operations from S1 to S4 for each frame of subsequent input images, splice the foreign object images of each frame, and synthesize the final three-dimensional road surface information and foreign object marking results.

Following the above technical solution, step S1 is specifically:

Input the infrared image of the laser line scan, and perform the grayscale value of each pixel in each column in turn. Step gradient operation, for each column of pixels, there is a maximum gradient value/> and minimum value/> , the number of rows that record the maximum gradient/> and the number of rows with the gradient minimum/> ,/> Indicates the upper boundary position of the laser line,/> Then it is the lower boundary position of the laser line, and the center position of the laser line is obtained/> :

Perform operations on the pixels in each column of the input image to extract the center point of the entire laser line;

Perform one-dimensional median filtering on the extracted y-coordinate of the center point of the laser line.

Following the above technical solution, step S2 is specifically:

of the extracted laser line Coordinates as forward neighborhood/> Ladder gradient operation,/> is an integer greater than or equal to 2;

Set an allowed fluctuation threshold , traverse the gradient value of each point of the extracted laser line , will be less than/> of/> Set to 0 to obtain a multi-step gradient curve of the processed laser line.

Following the above technical solution, in step S3, search for the minimum value on the multi-step gradient curve to the right. If the minimum value on the multi-step gradient curve is found, start looking for the maximum value. After finding the maximum value, these two points A foreign object pattern matching is formed, in which the minimum value represents the left boundary of the foreign object on the road surface, and the maximum value represents the right boundary of the foreign object on the road surface. The x-coordinates of these two points are expressed as and/> , and record;

Set a maximum allowed foreign body value , when/> , it is considered that there exists /> is the left border position,/> There is a foreign object at the right boundary position; if , it is considered that this pairing is incorrect, discard this pairing, and start from/> Starting from the position, step S3 is repeated.

Following the above technical solution, in step S4, the misidentified foreign objects are filtered by calculating the height information of the foreign objects, specifically as follows:

calculate to/> average height of laser lines/> :

;

Calculate the neighborhood height of its left and right neighbors ,/> ,in

calculate , set a height difference threshold , when satisfied/> or/> When , it is considered that this laser line is not a laser line formed by scanning foreign objects; otherwise, it is considered that this laser line is a laser line formed by scanning foreign objects.

The invention also provides a real-time detection system for foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching, including:

The laser line extraction module is used to input the infrared image of the road surface scanned by the infrared laser line and extract the laser lines in the infrared image of the road surface;

The multi-step gradient curve calculation module is used to perform multi-neighbor multi-step gradient calculation on the extracted y-coordinate of the laser line to obtain the multi-step gradient curve corresponding to the points on the laser line;

The characteristic pattern recognition module is used to find the minimum and maximum values on the multi-step gradient curve. According to the characteristics of the zero-crossing point of the image, the following characteristic patterns of foreign objects, pits and roadbed are defined:

On a multi-step gradient curve, if the gradient minimum appears first and then the gradient maximum appears within a certain window, this characteristic pattern represents foreign matter. The laser line between the gradient minimum and the gradient maximum The category label of the point on is set to Foreign Object;

If the gradient maximum value appears first, and then the gradient minimum value appears within a certain window range, then this feature mode is a pit, and the category label of the point on the laser line between the gradient maximum value and the gradient minimum value is set. for pits;

The remaining feature patterns are roadbed, and the category labels of the points on these laser lines are set to roadbed;

The inspection module is used to filter misidentified foreign objects by calculating the height information of the foreign objects, and obtain the position of the foreign objects based on the x-coordinate of the marked foreign objects;

The splicing module is used to splice multiple frames of foreign object marked images to synthesize the final three-dimensional road surface information and foreign object marking results.

Following the above technical solution, the multi-step gradient curve calculation module is specifically used for:

of the extracted laser line Coordinates as forward neighborhood/> Ladder gradient operation,/> is an integer greater than or equal to 2;

Set an allowed fluctuation threshold , traverse the gradient value of each point of the extracted laser line , will be less than/> of/> Set to 0 to obtain a multi-step gradient curve of the processed laser line.

Following the above technical solution, when the feature pattern recognition module performs foreign object pattern recognition, it specifically looks to the right for the minimum value on the multi-step gradient curve. If the minimum value on the multi-step gradient curve is found, it starts to look for the maximum value. After finding the maximum value, the two points form a foreign object pattern matching. The minimum value represents the left boundary of the foreign object on the road surface, and the maximum value represents the right boundary of the foreign object on the road surface. The x-coordinates of these two points are expressed as and/> , and record;

Set a maximum allowed foreign body value , when/> , it is considered that there exists /> is the left border position,/> There is a foreign object at the right boundary position; if , it is considered that this pairing is incorrect, discard this pairing, and start from/> Start position and recheck.

Continuing with the above technical solution, the inspection module filters misidentified foreign objects by calculating the height information of the foreign objects, specifically as follows:

calculate to/> average height of laser lines/> :

;

Calculate the neighborhood height of its left and right neighbors ,/> ,in

calculate , set a height difference threshold , when satisfied/> or/> When , it is considered that this laser line is not a laser line formed by scanning foreign objects; otherwise, it is considered that this laser line is a laser line formed by scanning foreign objects.

The present invention also provides a computer storage medium, which stores a computer program that can be executed by a processor. The computer program implements the real-time detection method of foreign objects on airport runway pavement by matching the image zero-crossing point characteristics and the foreign object feature pattern described in the above technical solution. .

The beneficial effects produced by the present invention are: The present invention defines the characteristic patterns of foreign objects, pits and roadbeds based on the extreme point matching rules on the multi-step gradient curve. This matching rule algorithm has low complexity and improves the efficiency of foreign object detection; Further verification of the height values between pairs of extreme points eliminates misjudged foreign objects and improves the accuracy of detection.

Description of the drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts.

Figure 1 is a flow chart of a real-time detection method for foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching according to an embodiment of the present invention;

Figure 2 is an input image according to the embodiment of the present invention;

Figure 3 is a center extraction image of the laser line in the embodiment of the present invention;

Figure 4 is a graph showing the forward neighborhood 4-step gradient calculation for the extracted laser lines;

Figure 5 is a multi-step gradient curve graph after filtering gradients below the threshold;

Figures 6(a)-6(c) are schematic diagrams of pattern matching of three foreign body features;

Figure 7 shows the foreign object marking map of a single frame image;

Figure 8 shows the results of two-dimensional foreign body marking;

Figure 9 is a three-dimensional foreign object point cloud result diagram;

Figure 10 is a physical map of the road surface.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention and are not intended to limit the present invention.

Example 1

As shown in Figure 1, this embodiment's real-time detection method of foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching mainly includes the following steps:

S1. Input the infrared image of the road surface scanned by the infrared laser line, and extract the laser lines in the infrared image of the road surface;

S2. Perform multi-neighbor multi-step gradient calculation on the extracted y-coordinate of the laser line to obtain the multi-step gradient curve corresponding to the points on the laser line;

S3. Find the minimum and maximum values on the multi-step gradient curve. These maximum and minimum values are actually the zero-crossing points of the image. The theoretical first-order derivative of the zero-crossing point of the image is the extreme value and the second-order derivative is zero. According to the above characteristics of the image zero-crossing point, the following characteristic patterns of foreign objects, pits and roadbed are defined:

On a multi-step gradient curve, if the gradient minimum appears first and then the gradient maximum appears within a certain window, this characteristic pattern represents foreign matter. The laser line between the gradient minimum and the gradient maximum The category label of the point on is set to Foreign Object;

If the gradient maximum value appears first, and then the gradient minimum value appears within a certain window range, then this feature mode is a pit, and the category label of the point on the laser line between the gradient maximum value and the gradient minimum value is set. for pits;

The remaining feature patterns are roadbed, and the category labels of the points on these laser lines are set to roadbed;

S4. Filter misidentified foreign objects by calculating the height information of the foreign objects, and obtain the position of the foreign objects based on the x-coordinate of the marked foreign objects;

S5: Repeat the operations from S1 to S4 for each frame of subsequent input images, splice the foreign object images of each frame, and synthesize the final three-dimensional road surface information and foreign object marking results.

This invention scans the infrared image of the road surface with a laser line, performs multi-neighbor multi-step gradient calculation on the extracted laser line y coordinate, obtains a multi-step gradient curve corresponding to the points on the laser line, and finds the minimum value and sum on the multi-step gradient curve. Maximum value, according to the characteristics of the zero-crossing point of the image, the following characteristic patterns of foreign objects, pits and roadbed are defined. This method has low complexity and can achieve efficient detection of foreign objects on the airport runway pavement, and filter misidentified foreign objects according to the height of the foreign objects, thus Improved detection accuracy.

Example 2

This embodiment is based on Embodiment 1, and specifically applies the method of Embodiment 1 to the detection of foreign objects on the road surface at a specific airport.

In this embodiment, the real-time detection method of foreign objects on airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching includes the following steps:

S1. Collect infrared images of the road surface scanned by infrared laser lines. The resolution of the images is . Input the infrared image of the laser line scan. The infrared image of the laser line is shown in Figure 2. The gray value of each pixel in each column is sequentially calculated. Ladder gradient operation (/> is an integer greater than or equal to 2, such as 2, 3, 4, 5...), in actual operation/> Take/> , the gradient template is .

Therefore, for each column of pixels, there is a maximum gradient value and minimum value , the number of rows that record the maximum gradient/> and the number of rows with the gradient minimum/> ,/> Indicates the upper boundary position of the laser line,/> is the lower boundary position of the laser line, so the center position of the laser line can be obtained/>

Perform the above operation on the pixels in each column of the input image to extract the center point of the entire laser line.

Finally, a one-dimensional median filter is performed on the y-coordinate of the extracted laser line center point. In a specific implementation, the window of the median filter is usually 3. The result of the extracted laser line is shown in Figure 3.

S2. There is a correlation between the multi-order differential gradient value of discrete points and the first-order derivative of continuous points. When there is a continuous function , and when using a set of equidistant discrete points to sample them to obtain discrete data points, the derivatives can be approximately calculated through differential gradients. /> The calculation method of the first-order differential gradient value is:

of the extracted laser lines Coordinates forward/> Step gradient calculation, usually for/> Take/> . y(i) represents the y-coordinate of the laser line in the i-th column, and y(j) represents the y-coordinate of the laser line in the j-th column. Therefore, for each extracted point of the laser line, there is a forward/> The calculated multi-step gradient curve is shown in Figure 4.

Then set an allowed fluctuation threshold , traverse the multi-step gradient value of each point of the extracted laser line/> , will be less than/> of/> Set to 0. As a result, a multi-step gradient curve of the processed laser line is obtained, as shown in Figure 5.

S3. Find the minimum and maximum values on the multi-step gradient curve. The method of finding extreme values is to find a gradient value whose absolute value is greater than the threshold on a multi-step gradient curve. After the point, perform extreme value judgment on this point: The judgment method of the maximum value point is: point, if satisfied,

Then, the first The point is the maximum value point. In the same way, if it satisfies

Then, the first The point is the minimum value point. Based on this, we find all the maximum and minimum points on the multi-step gradient curve.

According to the characteristics of the zero-crossing point of the image, the following characteristic patterns of foreign objects, pits and roadbed are defined:

On a multi-step gradient curve, if the gradient minimum appears first and then the gradient maximum appears within a certain window, this characteristic pattern represents foreign matter. The laser line between the gradient minimum and the gradient maximum The category label of the point on is set to foreign matter, as shown in Figure 6(a); if the gradient maximum value appears first, and then the gradient minimum value appears within a certain window range, then this feature mode is a pit, and the gradient maximum value appears. The category labels of the points on the laser line between the maximum value and the gradient minimum value are set as pits, as shown in Figure 6(b); the remaining feature patterns are roadbeds, and the category labels of the points on these laser lines are set Set as roadbed, as shown in Figure 6(c).

The specific matching method is to look for the minimum value on the multi-step gradient curve to the right. If the minimum value on the multi-step gradient curve is found, start looking for the maximum value. After the maximum value is found, the two points form a foreign body pattern. Match, where the minimum value represents the left boundary of the foreign object on the road surface, and the maximum value represents the right boundary of the foreign object on the road surface. The x-coordinates of these two points are expressed as and/> , and record it.

Since the roadbed may be uneven or have multiple foreign objects during the foreign object judgment process, in step S3, the minimum value on the multi-step gradient curve may be found to be paired with the next maximum value, resulting in incorrect pairing. Therefore, first set a maximum allowable foreign matter value ,/> The value of depends on the actual situation, usually // That’s it, when/>

It is considered that is the left border position,/> For the right boundary position a foreign body is possible if

It is considered that this pairing is incorrect, this pairing is discarded, and the Starting from the position, step S3 is repeated.

S4, for passing The judged extreme value pair needs to be further verified to see whether it is a foreign object. The reason is that when a depression appears on the road surface and the two depressions are close to each other, the roadbed between the two depressions may be misjudged as a foreign object. To address this issue, the verification method is:

calculate to/> average height of laser lines/>

Then calculate the average height of the left neighborhood and right neighborhood of this laser line ,/> ,in

calculate , set a height difference threshold , when satisfied/> or/> When , it is considered that this laser line is not a laser line formed by scanning foreign objects; otherwise, it is considered that this laser line is a laser line formed by scanning foreign objects.

According to the above process, we obtained the x-coordinates of several paired image zero-crossing points. Based on the x-coordinates of the zero-crossing points in the current image (that is, the obtained maximum and minimum values), the position of the foreign object can be obtained. Mark all laser line points between paired zero-crossing points as , indicating that these points are foreign objects, and other points are marked/> , indicating that these points are roadbed. The resolution of the input image is known from step S1, and a new image is created based on this resolution to mark the location of the foreign object. After traversing the above marks, the laser line point encounters the mark /> The current laser line point will be marked in red, and when it is encountered, it will be marked as/> are not marked. The foreign body marking results are shown in Figure 7. The first line is the input laser line infrared image, the second line is the extracted laser line image, and the foreign body is marked in red, and the third line is the gradient curve. The four lines are foreign body marks, of which the green line segment represents the roadbed and the red line segment represents foreign matter.

S5. Repeat the operations from S1 to S4 for each subsequent frame of image input. Since the position of the foreign object in each frame is obtained in step S4, combined with these position information, the foreign object images of each frame are spliced together to synthesize a The complete foreign body marking result chart and the final two-dimensional foreign body marking result are shown in Figure 8. In the previous steps, we also recorded the height information of the foreign object through the laser line information, based on which we can generate a three-dimensional point cloud result map of foreign object detection, as shown in Figure 9. The actual road map of the result map is shown in Figure 10.

Example 3

This embodiment is used to implement the above method embodiment. This embodiment is based on the airport runway pavement foreign object real-time detection system based on image zero-crossing point characteristics and foreign object feature pattern matching, including:

The laser line extraction module is used to input the infrared image of the road surface scanned by the infrared laser line and extract the laser lines in the infrared image of the road surface;

The multi-step gradient curve calculation module is used to perform multi-neighbor multi-step gradient calculation on the extracted y-coordinate of the laser line to obtain the multi-step gradient curve corresponding to the points on the laser line;

The characteristic pattern recognition module is used to find the minimum and maximum values on the multi-step gradient curve. According to the characteristics of the zero-crossing point of the image, the following characteristic patterns of foreign objects, pits and roadbed are defined:

On a multi-step gradient curve, if the gradient minimum appears first and then the gradient maximum appears within a certain window, this characteristic pattern represents foreign matter. The laser line between the gradient minimum and the gradient maximum The category label of the point on is set to Foreign Object;

If the gradient maximum value appears first, and then the gradient minimum value appears within a certain window range, then this feature mode is a pit, and the category label of the point on the laser line between the gradient maximum value and the gradient minimum value is set. for pits;

The remaining feature patterns are roadbed, and the category labels of the points on these laser lines are set to roadbed;

The inspection module is used to filter misidentified foreign objects by calculating the height information of the foreign objects, and obtain the position of the foreign objects based on the x-coordinate of the marked foreign objects;

The splicing module is used to splice multiple frames of foreign object marked images to synthesize the final three-dimensional road surface information and foreign object marking results.

Each module in the system of this embodiment is mainly used to implement each step of the above method embodiment, which will not be described again here.

This application also provides a computer-readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory Memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disks, optical disks, servers, App application malls, etc., on which computer programs and programs are stored The corresponding function is implemented when executed by the processor. When the computer-readable storage medium of this embodiment is executed by the processor, the real-time detection method of foreign objects on the airport runway pavement based on image zero-crossing point characteristics and foreign object feature pattern matching according to the method embodiment is implemented.

In summary, the airport road foreign object detection method of the present invention is real-time, efficient and accurate, and the processing speed can reach 300 frames per second. It can be applied to airports and other occasions where ground foreign objects detection is required, and improves the safety and work of aircraft and airports. efficiency and broad market application prospects.

It should be pointed out that according to the needs of implementation, each step/component described in this application can be split into more steps/components, or two or more steps/components or partial operations of steps/components can be combined into new ones. steps/components to achieve the objectives of the invention.

The sequence number of each step in the above embodiment does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the protection scope of the appended claims of the present invention.

Claims (10)

1. An airport runway pavement foreign matter real-time detection method based on image zero crossing point characteristic and foreign matter characteristic pattern matching is characterized by comprising the following steps:

s1, inputting an infrared laser line scanned road surface infrared image, and extracting laser lines in the road surface infrared image;

s2, performing multi-neighborhood multi-order gradient calculation on the y coordinate of the extracted laser line to obtain a multi-order gradient curve corresponding to a point on the laser line;

s3, finding minimum values and maximum values on the multi-order gradient curve, and defining the following characteristic modes of foreign matters, pits and roadbed according to the characteristics of the image zero crossing points:

on a multi-order gradient curve, if a gradient minimum value appears firstly and then a gradient maximum value appears in a certain window range, the characteristic mode represents foreign matters, and a category label from the gradient minimum value to a point on a laser line between the gradient maximum values is set as the foreign matters;

if the gradient maximum value appears firstly and then the gradient minimum value appears in a certain window range, the characteristic mode is a pit, and the category label of the point on the laser line from the gradient maximum value to the gradient minimum value is set as the pit;

the rest characteristic modes are roadbed, and category labels of points on the laser lines are set as roadbed;

s4, filtering misjudged foreign matters by calculating the height information of the foreign matters, when the absolute value of the difference value between the average height of the foreign matters and the average height of the left neighborhood or the right neighborhood of the foreign matters is smaller than a preset height difference threshold value, considering the foreign matters as misjudged foreign matters, otherwise, considering the foreign matters as foreign matters, and obtaining the positions of the foreign matters according to the x coordinates of marked foreign matters;

s5, repeating the operations from S1 to S4 on each frame of image which is input subsequently, splicing the foreign object images of each frame, and synthesizing the final pavement three-dimensional information and the foreign object marking result.

2. The method for detecting the foreign objects on the airport runway pavement in real time based on the matching of the image zero crossing characteristic and the foreign object characteristic mode according to claim 1, wherein the step S1 is specifically as follows:

inputting an infrared image scanned by a laser line, and sequentially carrying out gray value of each pixel point of each columnThe step gradient operation has a maximum value of gradient value for each column of pixel points>And minimum->Number of rows recording gradient maximum +.>And the number of lines of gradient minima->，/>Represents the upper boundary position of the laser line, +.>Then the lower boundary position of the laser line, the central position of the laser line is obtained +.>：

Calculating pixel points of each column of the input image, and extracting a central point of the whole laser line;

and carrying out one-dimensional median filtering on the y coordinate of the extracted laser line central point.

3. The method for detecting the foreign objects on the airport runway pavement in real time based on the matching of the image zero crossing characteristic and the foreign object characteristic mode according to claim 1, wherein the step S2 is specifically:

for extracted laser linesCoordinate forward neighborhood->Ladder degree calculation (I/O)>Is an integer of 2 or more;

setting an allowable fluctuation thresholdGradient value of each point of traversing one-pass extracted laser line +.>Will be less than->Is->Set to 0, a multi-step gradient profile of the processed laser line is obtained.

4. According to claimThe method for detecting the foreign matter on the airport runway pavement based on the matching of the image zero crossing point characteristic and the foreign matter characteristic mode is characterized in that in the step S3, the minimum value on the multi-step gradient curve is searched to the right, if the minimum value on the multi-step gradient curve is found, the maximum value is searched, the two points form the foreign matter mode matching after the maximum value is found, wherein the minimum value represents the left boundary of the foreign matter on the pavement, the maximum value represents the right boundary of the foreign matter on the pavement, and the x coordinates of the two points are represented asAnd->And recording;

setting a maximum allowable foreign matter valueWhen->Then consider to exist toLeft boundary position->Foreign matter exists for the right boundary position; if->Then the pairing is considered incorrect, the pairing is discarded and the pairing is taken from +.>The position starts and step S3 is repeatedly performed.

5. The method for detecting the foreign matter on the runway pavement in real time according to the matching of the image zero crossing characteristic and the foreign matter characteristic pattern of claim 4, wherein the step S4 is characterized in that the filtering misjudged foreign matter by calculating the height information of the foreign matter is specifically as follows:

calculation ofTo->Average height of laser lines in between +.>：

；

Calculating the neighborhood height of the left neighborhood and the right neighborhood、/>Wherein

Calculation ofSetting a height difference threshold +.>When meeting->Or->If the laser line is not a laser line scanned to form a foreign object, the laser line is considered to be a laser line scanned to form a foreign object.

6. An airport runway pavement foreign matter real-time detection system based on image zero crossing characteristic and foreign matter characteristic pattern matching, which is characterized by comprising:

the laser line extraction module is used for inputting the road surface infrared image scanned by the infrared laser lines and extracting the laser lines in the road surface infrared image;

the multi-order gradient curve calculation module is used for performing multi-neighborhood multi-order gradient calculation on the y coordinate of the extracted laser line to obtain a multi-order gradient curve corresponding to a point on the laser line;

the characteristic pattern recognition module is used for finding minimum values and maximum values on the multi-order gradient curve, and defining the following characteristic patterns of foreign matters, pits and roadbeds according to the characteristics of the image zero crossing points:

on a multi-order gradient curve, if a gradient minimum value appears firstly and then a gradient maximum value appears in a certain window range, the characteristic mode represents foreign matters, and a category label from the gradient minimum value to a point on a laser line between the gradient maximum values is set as the foreign matters;

if the gradient maximum value appears firstly and then the gradient minimum value appears in a certain window range, the characteristic mode is a pit, and the category label of the point on the laser line from the gradient maximum value to the gradient minimum value is set as the pit;

the rest characteristic modes are roadbed, and category labels of points on the laser lines are set as roadbed;

the detection module is used for filtering the misjudged foreign matters by calculating the height information of the foreign matters, when the absolute value of the difference value between the average height of the foreign matters and the average height of the left neighborhood or the right neighborhood of the foreign matters is smaller than a preset height difference threshold value, the foreign matters are considered as misjudged foreign matters, otherwise, the foreign matters are considered as foreign matters, and the positions of the foreign matters are obtained according to the x coordinates of the marked foreign matters;

and the splicing module is used for splicing the multi-frame images marked by the foreign matters to synthesize the final pavement three-dimensional information and the foreign matter marking result.

7. The airport runway pavement foreign object detection system based on image zero crossing characteristic and foreign object characteristic pattern matching of claim 6, wherein the multi-order gradient curve calculation module is specifically configured to:

for extracted laser linesCoordinate forward neighborhood->Ladder degree calculation (I/O)>Is an integer of 2 or more;

setting an allowable fluctuation thresholdGradient value of each point of traversing one-pass extracted laser line +.>Will be less than->Is->Set to 0, a multi-step gradient profile of the processed laser line is obtained.

8. The system for detecting the foreign object on the runway pavement in real time based on the matching of the characteristic of the image zero crossing point and the characteristic pattern of the foreign object according to claim 6, wherein the characteristic pattern recognition module searches for the minimum value on the multi-step degree curve to the right specifically when the characteristic pattern recognition is performed, and starts if the minimum value on the multi-step degree curve is foundSearching a maximum value, and forming foreign matter pattern matching by the two points after the maximum value is found, wherein the minimum value represents the left boundary of foreign matter on the road surface, the maximum value represents the right boundary of foreign matter on the road surface, and the x coordinates of the two points are represented asAnd->And recording;

setting a maximum allowable foreign matter valueWhen->Then consider to exist toLeft boundary position->Foreign matter exists for the right boundary position; if->Then the pairing is considered incorrect, the pairing is discarded and the pairing is taken from +.>The position starts and is re-detected.

9. The airport runway pavement foreign matter real-time detection system based on image zero crossing characteristic and foreign matter characteristic pattern matching of claim 8, wherein the foreign matter misjudged by calculating the height information of the foreign matter in the inspection module is specifically:

calculation ofTo->Average height of laser lines in between +.>：

；

Calculating the neighborhood height of the left neighborhood and the right neighborhood、/>Wherein

Calculation ofSetting a height difference threshold +.>When meeting->Or->If the laser line is not a laser line scanned to form a foreign object, the laser line is considered to be a laser line scanned to form a foreign object.

10. A computer storage medium in which a computer program executable by a processor is stored, the computer program implementing the method for detecting an airport runway surface foreign matter in real time by matching the image zero-crossing characteristic and the foreign matter characteristic pattern according to any one of claims 1 to 5.