CN110033490B - Airport low-slow small target prevention and control method based on photoelectric image automatic identification - Google Patents

Abstract

The invention discloses an airport low-slow and small target prevention and control method based on photoelectric image automatic identification. image; perform feature extraction on each image, classify the target type corresponding to the image according to the extracted image features, and obtain a low-slow and small-target classifier; preprocess the target image for prevention and control, extract the image features, and then use the The feature is input to the classifier to obtain the type of the target to be prevented and controlled; the location of the target to be prevented and controlled is determined, and the corresponding prevention and control equipment is driven to work in combination with the type of the target to be prevented and controlled. The invention realizes multiple prevention and control strategies in the airport area by classifying different targets, realizes the optimal combination of different low-slow and small prevention and control equipment, greatly reduces the impact of low-slow and small targets on the airport airspace, and overcomes the drawbacks of traditional manual monitoring of airspace and a single control strategy Limitations, improve the efficiency and accuracy of the prevention and control of low-slow and small targets.

Description

A method for preventing and controlling low, slow and small targets in airports based on automatic photoelectric image recognition

technical field

The invention belongs to the technical field of intelligent low-slow small target prevention and control, in particular to an airport low-slow and small target prevention and control method based on photoelectric image automatic identification.

Background technique

"Low, slow and small" targets refer to aircraft and floating objects that fly at low and ultra-low altitudes, are slow, have small volumes, and are not easily detected. It mainly includes light and ultra-light aircraft, light helicopters, gliders, delta wings, powered delta wings, powered parachutes, paragliders, hot air balloons, airships, drones, aviation models, aerospace models, air balloons, tethered balloons, etc. These objects are low-cost, simple to control, easy to carry, and easy to obtain. They also have strong liftoff and are difficult to find and dispose of. They are easily used as tools for carrying explosives, dropping biochemical agents, and disseminating leaflets, seriously threatening the safety of major events and key areas. Safeguard jobs.

Bird strike refers to the accident caused by the collision between the aircraft running at high speed in the air or on the ground and the birds in the air. During the impact, the bird crash is more similar to a soft body crash because the huge change in the movement state of the bird exceeds the ability of the bird's body to bear, causing the bird to die and its limbs to split. Bird strikes have the characteristics of suddenness and variability. Flight accidents caused by bird strikes will not only bring economic losses, but also bring greater hidden dangers to life and safety, threatening the personal safety of passengers and flight crews at all times. Bird strikes can cause damage to aircraft components, and severe bird strikes can lead to aircraft crashes.

Traditional airport airspace detection usually adopts radar detection, optical detection, radio signal detection and so on. However, due to the small radar reflection area of low-slow small targets and birds, and the Doppler effect caused by the slow flight speed is not obvious, it is easy to cause target misjudgment. In addition, optical detection is mainly composed of infrared automatic search technology, image processing technology and high-precision turntable control technology. Existing low, slow, small and bird detection is becoming more and more mature, including the American Merlin radar system, the Canadian Goshawk radar system, the Beijing Zhongke Robin radar system, and the bird detection radar system of the leading general aviation airport in Wuhan, etc. Some of them have achieved detection and removal linkage, but did not propose classified prevention and control strategies and methods.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide an airport low-slow and small target prevention and control method that can overcome the disadvantages of traditional manual monitoring of airspace and the limitation of a single control strategy, and improve the control efficiency and accuracy of low-slow and small targets.

The technical solution for realizing the purpose of the present invention is: an airport low-slow and small-target prevention and control method based on automatic identification of photoelectric images, comprising the following steps:

Step 1. Obtain multiple images of each low-slow and small target from the network, and conduct field measurements to collect multiple images of each low-slow and small target;

Step 2, perform feature extraction on each image obtained in step 1, classify the target type corresponding to the image according to the extracted image features, and obtain a low-slow and small-target classifier;

Step 3. Preprocess the image of the target to be prevented and controlled, and extract the image features, and then input the feature to the classifier to obtain the type of the target to be prevented and controlled;

Step 4: Determine the position of the target to be prevented and controlled, and drive the corresponding prevention and control equipment to work in combination with the type of the target to be prevented and controlled obtained in step 3.

Compared with the prior art, the present invention has the following significant advantages: 1) different prevention and control devices are driven according to different types of low-slow and small targets, thereby improving the automation level of low-slow and small target control; 2) by extracting image features The method realizes the automatic identification and prevention of target types, and realizes the linkage and integration of detection and prevention and control, which can reduce labor costs; 3) Expand the training set through web crawler technology to obtain more training images, thereby improving the classifier model. 4) Through a series of preprocessing such as denoising, smoothing, and background segmentation on the target image, the influence of noise on the results is reduced, and the accuracy of prevention and control is improved; 5) By predicting the position of the target at the next moment, driving the prevention and control The control equipment is launched to the future position of the target to improve the hit rate.

The present invention will be described in further detail below in conjunction with the accompanying drawings.

Description of drawings

FIG. 1 is a flow chart of a method for preventing and controlling low, slow, and small targets in an airport based on the automatic identification of photoelectric images according to the present invention.

FIG. 2 is a grayscale image of an image of a target to be controlled in an embodiment of the present invention.

FIG. 3 is an image obtained by filtering FIG. 2 in an embodiment of the present invention.

FIG. 4 is a diagram of a result of target type discrimination in an embodiment of the present invention.

FIG. 5 is a schematic diagram of spatial coordinates established in an embodiment of the present invention.

FIG. 6 is a flow chart of the prevention and control processing after the target is detected in the present invention.

Detailed ways

1 , a method for preventing and controlling low, slow and small targets in an airport based on automatic photoelectric image recognition of the present invention includes the following steps:

Step 1. Obtain multiple images of each low-slow and small target from the network, and conduct field measurements to collect multiple images of each low-slow and small target;

Step 2, perform feature extraction on each image obtained in step 1, classify the target type corresponding to the image according to the extracted image features, and obtain a low-slow and small-target classifier;

Step 3. Preprocess the image of the target to be prevented and controlled, and extract the image features, and then input the feature to the classifier to obtain the type of the target to be prevented and controlled;

Step 4: Determine the position of the target to be prevented and controlled, and drive the corresponding prevention and control equipment to work in combination with the type of the target to be prevented and controlled obtained in step 3.

Further preferably, the low-slow and small targets include drones, balloons, and birds.

Further preferably, in step 1, multiple images of each low-slow and small target are obtained from the network, specifically: crawling multiple images of each low-slow and small target from the network through a web crawler.

Further preferably, in step 1, multiple images of each low-slow and small target are obtained from the network, specifically: using the name of the low-slow and small target as a keyword, and crawling each low-slow and small target from the network through a Python crawler. image.

Further preferably, in step 2, feature extraction is performed on each image obtained in step 1, specifically: using the Haar feature extraction method to extract image features.

Further, before the feature extraction in step 2, it also includes preprocessing each image obtained in step 1, and the preprocessing includes: denoising, smoothing, and background segmentation.

Further preferably, in step 2, the target type corresponding to the image is classified according to the extracted image features, and specifically, the support vector machine SVM is used for classification.

Further, determining the position of the target to be prevented and controlled as described in step 4, combined with the type of the target to be prevented and controlled obtained in step 3, to drive the corresponding prevention and control equipment to work, specifically:

Step 4-1. Establish a spatial coordinate system for the airport area, where the airport area includes the runway and the area along both sides of the runway; place the photoelectric detection equipment at any position on both sides of the airport runway, take this position as the origin O, and move along the plane along the runway. The direction of the runway is the y-axis, the horizontal direction perpendicular to the y-axis is the x-axis, and the direction perpendicular to the xOy plane pointing to the sky is the z-axis; thus determine the position of the current position of the target to be controlled in the space coordinate system;

Step 4-2: Determine the corresponding prevention and control equipment according to the type of the target to be prevented and controlled, wherein the corresponding relationship between the type of the target to be prevented and controlled and the prevention and control equipment is defined by the user;

Step 4-3, determining the position of the prevention and control equipment corresponding to the target to be prevented and controlled in the space coordinate system;

Step 4-4, the photoelectric detection device obtains the current position and speed of the target to be prevented and controlled, thereby estimating the position of the target to be prevented and controlled in the space coordinate system at the next moment;

Step 4-5, move the origin O of the space coordinate system to the location of the prevention and control equipment, obtain a new space coordinate system, and convert the position of the target to be controlled in the original space coordinate system at the next moment to the new space coordinate system, After that, it is converted to the spherical coordinate system with the prevention and control equipment as the origin, and the prevention and control equipment interferes with the prevention and control target according to the converted position.

Further preferably, the corresponding relationship between the target type to be prevented and controlled and the prevention and control equipment described in step 4-2 is defined by the user, specifically: if the target type to be prevented and controlled is an unmanned aerial vehicle, the corresponding prevention and control equipment is an unmanned aerial vehicle. Interference equipment; if the target type to be prevented and controlled is balloon, the corresponding prevention and control equipment is laser equipment; if the type of target to be prevented and controlled is birds, the corresponding prevention and control equipment is bird repelling equipment.

Further preferably, the estimation of the position of the target to be prevented and controlled in the space coordinate system at the next moment in step 4-4 specifically adopts a least squares filtering algorithm.

The present invention will be described in further detail below in conjunction with the embodiments.

Example

1, the present invention based on the photoelectric image automatic identification of the airport low-slow and small target prevention and control method, including the following content:

1. Obtain multiple images of each low-slow and small target from the network through a web crawler, and conduct on-site measurements to obtain multiple images of each low-slow and small target.

2. Use the Haar feature extraction method to perform feature extraction on each image obtained in the above 1, and classify the target type corresponding to the image according to the extracted image features to obtain a low-slow and small-target classifier.

3. In this embodiment, the target image to be controlled and controlled is set to 128*128 pixels, and then converted into a grayscale image as shown in Figure 2, and then median filtering is performed on the grayscale image, and the filtering result is shown in Figure 3. Show.

4. Preprocess the image of the target to be prevented and controlled, and extract the image features, and then input the feature into the type of the target to be prevented and controlled obtained by the low-slow and small target classifier. In this embodiment, the type of the target to be prevented and controlled obtained is None man-machine, as shown in Figure 4.

5. With reference to Figure 5, establish a spatial coordinate system for the airport area, where the airport area includes the runway and the area along both sides of the runway; place the photoelectric detection equipment at any position on both sides of the airport runway, and take this position as the origin O, along the runway. The direction of the runway is the y-axis, the horizontal direction perpendicular to the y-axis is the x-axis, and the direction perpendicular to the xOy plane pointing to the sky is the z-axis; thus, the current position of the target to be controlled is determined in the space coordinate system.

6. Referring to FIG. 6, when different target types are detected, different prevention and control equipment needs to be selected. In this embodiment, the prevention and control equipment corresponding to the type of the target to be prevented and controlled is a drone interference equipment.

7. Obtain the position of the drone interference device in the space coordinate system. In this embodiment, the position coordinate of the drone interference device is [0 1000 0] ^T .

8. The photoelectric detection equipment obtains the current position [100 2000 50] ^T and speed [5 15-8] ^T m/s of the target to be prevented and controlled, thereby predicting the position of the target to be prevented and controlled in the space coordinate system at the next moment is [105 2015 42] ^T .

高低角

无人机干扰设备指向该方向进行发射干扰。9. Translate the above coordinate system so that the origin of the coordinates is the location of the UAV interference device, and the device running clock step is defined as 1 second, then the position of the target future point in the new coordinate system is [105 1015 42] ^T , convert the position to spherical coordinate system as azimuth

high and low angle

The drone jamming device points in this direction to transmit jamming.

The invention takes the acquired multiple images as the basis, acquires the target image features to train the target type classifier, then discriminates the target type according to the actually detected target image, and then places the target in the detection coordinate system according to the established space rectangular coordinate system. The lower position and future point position are converted to the coordinate system of the prevention and control equipment, and the prevention and control equipment adjusts the launch direction according to the position, which reduces the error of manual operation, realizes the integration of detection and prevention and control, and greatly reduces the labor cost.

Claims (9)

1. A method for preventing and controlling low and slow small targets of an airport based on photoelectric image automatic identification is characterized by comprising the following steps:

step 1, acquiring a plurality of images of each low-slow small target from a network, and carrying out field actual measurement to acquire a plurality of images of each low-slow small target;

step 2, extracting the characteristics of each image obtained in the step 1, and classifying the target types corresponding to the images according to the extracted image characteristics to obtain a low-slow small target classifier;

step 3, preprocessing the image of the target to be prevented and controlled, extracting image characteristics, and then inputting the characteristics into the classifier to acquire the type of the target to be prevented and controlled;

step 4, determining the position of the target to be prevented and controlled, and driving corresponding prevention and control equipment to work by combining the type of the target to be prevented and controlled obtained in the step 3; the method specifically comprises the following steps:

step 4-1, establishing a space coordinate system for an airport area, wherein the airport area comprises a runway and areas along two sides of the runway; placing photoelectric detection equipment at any position of two side areas of an airport runway, taking the position as an original point O, taking the direction along the airport runway as a y axis, taking the horizontal direction vertical to the y axis as an x axis, and taking the direction vertical to the xOy plane and pointing to the sky as a z axis; thereby determining the position of the current position of the target to be controlled in the space coordinate system;

step 4-2, determining corresponding prevention and control equipment according to the type of the target to be prevented and controlled, wherein the corresponding relation between the type of the target to be prevented and controlled and the prevention and control equipment is customized by a user;

4-3, determining the position of the prevention and control equipment corresponding to the target to be prevented and controlled in the space coordinate system;

4-4, acquiring the current position and the speed of the target to be prevented and controlled by the photoelectric detection equipment, and estimating the position of the target to be prevented and controlled in the space coordinate system at the next moment;

and 4-5, moving the original point O of the space coordinate system to the position of the prevention and control equipment to obtain a new space coordinate system, converting the position of the target to be prevented and controlled in the original space coordinate system at the next moment into the new space coordinate system, then converting the position into a spherical coordinate system with the prevention and control equipment as the original point, and interfering the target to be prevented and controlled by the equipment to be prevented and controlled according to the converted position.

2. The airport low and slow small target prevention and control method based on automatic photoelectric image recognition is characterized in that the low and slow small targets comprise unmanned planes, balloons and birds.

3. The airport low-slow small target prevention and control method based on photoelectric image automatic identification as claimed in claim 1 or 2, wherein step 1 is to obtain a plurality of images of each low-slow small target from the network, specifically: and crawling a plurality of images of each low-slow small target from the network through a web crawler.

4. The airport low-slow small target prevention and control method based on photoelectric image automatic identification as claimed in claim 3, wherein step 1 is to obtain a plurality of images of each low-slow small target from the network, specifically: and (4) crawling a plurality of images of each low-slow small target from the network by a Python crawler by taking the name of the low-slow small target as a keyword.

5. The airport low-slow small target prevention and control method based on photoelectric image automatic identification as claimed in claim 4, wherein step 2 performs feature extraction on each image obtained in step 1, specifically: and extracting image features by adopting a Haar feature extraction method.

6. The airport low-slow small target prevention and control method based on automatic photoelectric image recognition as claimed in claim 5, wherein before the feature extraction in step 2, each image obtained in step 1 is preprocessed, and the preprocessing comprises: denoising, smoothing and background segmentation.

7. The airport low-slow small target prevention and control method based on photoelectric image automatic identification as claimed in claim 6, wherein in step 2, the target types corresponding to the images are classified according to the extracted image features, specifically by using a Support Vector Machine (SVM).

8. The airport low-slow small target prevention and control method based on photoelectric image automatic identification as claimed in claim 1, wherein the corresponding relationship between the type of target to be prevented and controlled and the prevention and control device in step 4-2 is customized by a user, specifically: if the type of the target to be prevented and controlled is an unmanned aerial vehicle, the corresponding prevention and control equipment is unmanned aerial vehicle interference equipment; if the type of the target to be prevented and controlled is a balloon, the corresponding prevention and control equipment is laser equipment; if the type of the target to be prevented and controlled is birds, the corresponding prevention and control equipment is bird repelling equipment.

9. The airport low-slow small target prevention and control method based on automatic photoelectric image identification as claimed in claim 1, wherein in step 4-4, the position of the target to be prevented and controlled in the space coordinate system at the next moment is estimated, specifically using a least square filtering algorithm.

Images