CN106951823A - A kind of infrared image undercarriage automatic identifying method - Google Patents

Abstract

The invention discloses a kind of infrared image undercarriage automatic identifying method, comprise the following steps：Step 1, input infrared image I；Step 2, statistics infrared image I histogram His, calculates the segmentation threshold Th of foreground and background；Step 3, using segmentation threshold Th by infrared image I binaryzations, binary map Bi Img are obtained；Step 4, the position Pos of two engines of aircraft is positioned₁And Pos₂；Step 5, undercarriage distributed areas before and after calculating, respectively Area₁、Area₂And Area₃；Step 6, fall into a trap calculation undercarriage characteristic parameter F in distributed areas₁、F₂And F₃；Step 7, according to characteristic parameter F₁、F₂And F₃Complete undercarriage identification.Undercarriage automatic identification technology recognition correct rate of the present invention is high, operand is small, be easy to hardware real-time implementation.

Description

A kind of infrared image undercarriage automatic identifying method

Technical field

The present invention relates to a kind of infrared image undercarriage automatic identifying method.

Background technology

The abundant opening of undercarriage and the prerequisite that locking is aircraft safety landing.At present, although military-civil aircraft is all The machinery, light prompt mark and instrument for being provided with undercarriage indicate that part aircraft also has images first-class monitoring out of my cabin Equipment can observe undercarriage.But observer, manual confirmation need to be set up on ground toward contact during aircraft landing The open mode of undercarriage, to ensure to be perfectly safe.And manual observation is many due to being limited by weather, period, intensity of illumination etc. It is winged to meet that the configured information of undercarriage can not may in time, be effectively provided under the influence of extraneous factor, extreme case The demand of quick decision before machine landing, so being helped in infrared electro, that the undercarriage based on infrared image is added in drop system is full-automatic Identification technology just has very important significance.

In infrared image target detection, identification field, domestic and foreign scholars have carried out numerous studies, and obtain many researchs Achievement, main thought is according to the marginal information of target in image, target local gray level average and variance, target speed etc. The characteristic information of itself opens target with background segment.Undercarriage can not simply use traditional object detection method, lead to Cross foreground segmentation and directly distinguish in undercarriage region from aircraft overall structure and analyzed.Also need to solve following key Problem：(1) aircraft yardstick and attitude are accurately positioned under uncertain visual field, the structure of aircraft is extracted；(2) aircraft ring frames are based on Appropriate adaptivenon-uniform sampling, positioning are carried out to undercarriage region and is extracted.(3) rack-like of rising and falling is carried out to the undercarriage region of extraction The automatic identification of state；(4) most existing target identification method operands are big, are not easy to hardware real-time implementation.

The content of the invention

The purpose of the present invention be for double aircrafts held up of carrying out the coffin upon burial, design a kind of simple method, strong applicability, work well, And it is adapted to the infrared image undercarriage automatic identification technology of hardware real-time implementation.

The implementation steps of technical solution of the present invention are as follows：

Step 1, input infrared image I；

Step 2, statistics infrared image I histogram His, calculates the segmentation threshold Th of foreground and background；

Step 3, using segmentation threshold Th by infrared image I binaryzations, binary map Bi Img are obtained；

Step 4, the position Pos of two engines of aircraft is positioned₁And Pos₂；

Step 5, undercarriage distributed areas before and after calculating, respectively Area₁、Area₂And Area₃；

Step 6, fall into a trap calculation undercarriage characteristic parameter F in distributed areas₁、F₂And F₃；

Step 7, according to characteristic parameter F₁、F₂And F₃Complete undercarriage identification.

Step 2 comprises the following steps：

Step 2-1, calculates infrared image I histogram His (k) { k=0 ..., 255 }, k represents gray level；

Step 2-2, setting thresholding T₁, 1 × N is carried out from high grade grey level to low gray level to histogram His and draws window processing, if Draw the pixel count His (k) of current gray level in window>T₁, current gray level level is judged as background value, otherwise, it is determined that current gray level level is Prospect value, wherein, 5≤N≤15, N represents window width, 100≤T₁≤150；

Step 2-3,1 × N of statistics draws His (k) in window>T₁Number Count, if Count/N>0.8, then by the k at N/2 Value is determined as segmentation threshold Th, otherwise, continues to carry out 1 × N strokes of window processing from high grade grey level to low gray level, until finding segmentation Untill threshold value.

Step 4 includes：

Infrared imaging gray value depends on the difference of target and background, and difference is bigger, and gray value is higher.Aircraft IR signature Show as engine gray value highest.To avoid blind element, noise etc. from influenceing, from high grade grey level to low gray scale in histogram His Level is searched, will pixel count His (k) first>40 gray level is determined as engine gray value, according to the gray value in infrared image The position Pos of middle location engine₁And Pos₂, their coordinate is respectively (x₁,y₁) and (x₂,y₂), x₁,y₁Position is represented respectively Pos₁Abscissa and ordinate, x₂,y₂Position Pos is represented respectively₂Abscissa and ordinate, k represents gray level.

Step 5 includes：

Undercarriage distribution characteristics is：Nose-gear region Area₁Near the axis of two engines, rear undercarriage Region Area₂And Area₃It is located at the position Pos of two engines respectively₁And Pos₂Near, with two engine location coordinate (x₁, y₁) and (x₂,y₂) it is reference point, x₁<x₂, rectangular coordinate system is set up, then：

Area₁It is expressed as：Original position in transverse axis x directions isStop bits in transverse axis x directions It is set toIt is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Area₂It is expressed as：Original position in transverse axis x directions is x₁；End position in transverse axis x directions isIt is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Area₃It is expressed as：Original position in transverse axis x directions isEnd position in transverse axis x directions is x₂；It is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Wherein, T₂Represent length-width ratio, 2≤T₂≤4。

Step 6 includes：

In binary map Bi Img, it is assumed that front and rear undercarriage distributed areas Area₁、Area₂And Area₃Pixel value is 1 y Direction maximum position is respectively y_max1、y_max2And y_max3, then characteristic parameter F₁、F₂And F₃Expression formula is as follows：

Step 7 includes：

If F₁>T₃, then judge that nose-gear is in down state, otherwise, it is determined that nose-gear is in collapsed state；

If F₂>T₃, then judge that rear left undercarriage is in down state, otherwise, it is determined that rear left undercarriage is in collapsed state；

If F₃>T₃, then right landing gear is in down state after judging, otherwise, it is determined that rear right landing gear is in collapsed state；

Wherein, 0.5≤T₃≤1.Judge T according to actual conditions₁、T₂、T₃Specific value.

Beneficial effect：The present invention compared with prior art, with following remarkable advantage：(1) using based on statistics with histogram Adaptive targets dividing method, do not influenceed by aircraft yardstick and attitude, being capable of Accurate Segmentation target；(2) risen and fallen by setting up Frame identification model, to the accurate identification of undercarriage on the premise of pixel count is few；(3) from aircraft engine as reference, to rising Fall frame positioning, undercarriage recognition correct rate can be improved；(4) high exponent arithmetic(al) and labyrinth is not present, algorithm operation quantity is small, it is easy to Hardware real-time implementation.

Brief description of the drawings

The present invention is done with reference to the accompanying drawings and detailed description and further illustrated, it is of the invention above-mentioned or Otherwise advantage will become apparent.

Fig. 1 is the flow chart of infrared image undercarriage automatic identification technology of the present invention.

Fig. 2 a are original images.

Fig. 2 b are binary maps.

Fig. 2 c are remote undercarriage recognition results.

Fig. 3 a are original images.

Fig. 3 b are binary maps.

Fig. 3 c are closely undercarriage recognition effect figures.

Embodiment

Below in conjunction with the accompanying drawings and embodiment the present invention will be further described.

The implementation steps of technical solution of the present invention are as follows：

Step 1, input infrared image I；

Step 2, statistics infrared image I histogram His, calculates the segmentation threshold Th of foreground and background；

Step 3, using segmentation threshold Th by infrared image I binaryzations, binary map Bi Img are obtained；

Step 4, the position Pos of two engines of aircraft is positioned₁And Pos₂；

Step 5, undercarriage distributed areas before and after calculating, respectively Area₁、Area₂And Area₃；

Step 6, fall into a trap calculation undercarriage characteristic parameter F in distributed areas₁、F₂And F₃；

Step 7, according to characteristic parameter F₁、F₂And F₃Complete undercarriage identification.

Step 2 comprises the following steps：

Step 2-1, calculates infrared image I histogram His (k) { k=0 ..., 255 }, k represents gray level；

Step 2-2, setting thresholding T₁, 1 × N is carried out from high grade grey level to low gray level to histogram His and draws window processing, if Draw the pixel count His (k) of current gray level in window>T₁, current gray level level is judged as background value, otherwise, it is determined that current gray level level is Prospect value, wherein, 5≤N≤15, N represents window width, 100≤T₁≤150；

Step 2-3,1 × N of statistics draws His (k) in window>T₁Number Count, if Count/N>0.8, then by the k at N/2 Value is determined as segmentation threshold Th, otherwise, continues to carry out 1 × N strokes of window processing from high grade grey level to low gray level, until finding segmentation Untill threshold value.

Step 4 includes：

Infrared imaging gray value depends on the difference of target and background, and difference is bigger, and gray value is higher.Aircraft IR signature Show as engine gray value highest.To avoid blind element, noise etc. from influenceing, from high grade grey level to low gray scale in histogram His Level is searched, will pixel count His (k) first>40 gray level is determined as engine gray value, according to the gray value in infrared image The position Pos of middle location engine₁And Pos₂, their coordinate is respectively (x₁,y₁) and (x₂,y₂), x₁,y₁Position is represented respectively Pos₁Abscissa and ordinate, x₂,y₂Position Pos is represented respectively₂Abscissa and ordinate, k represents gray level.

Step 5 includes：

Undercarriage distribution characteristics is：Nose-gear region Area₁Near the axis of two engines, rear undercarriage Region Area₂And Area₃It is located at the position Pos of two engines respectively₁And Pos₂Near, with two engine location coordinate (x₁, y₁) and (x₂,y₂) it is reference point, x₁<x₂, rectangular coordinate system is set up, then：

Area₁It is expressed as：Original position in transverse axis x directions isStop bits in transverse axis x directions It is set toIt is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Area₂It is expressed as：Original position in transverse axis x directions is x₁；End position in transverse axis x directions isIt is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Area₃It is expressed as：Original position in transverse axis x directions isEnd position in transverse axis x directions is x₂；It is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Wherein, T₂Represent length-width ratio, 2≤T₂≤4。

Step 6 includes：

In binary map Bi Img, it is assumed that front and rear undercarriage distributed areas Area₁、Area₂And Area₃Pixel value is 1 y Direction maximum position is respectively y_max1、y_max2And y_max3, then characteristic parameter F₁、F₂And F₃Expression formula is as follows：

Step 7 includes：

If F₁>T₃, then judge that nose-gear is in down state, otherwise, it is determined that nose-gear is in collapsed state；

If F₂>T₃, then judge that rear left undercarriage is in down state, otherwise, it is determined that rear left undercarriage is in collapsed state；

If F₃>T₃, then right landing gear is in down state after judging, otherwise, it is determined that rear right landing gear is in collapsed state；

Wherein, 0.5≤T₃≤1。

Embodiment

Thermal infrared imager focal plane arrays (FPA) size is 640 × 512, and working frame frequency is 50 frame per second.Image processing platform is adopted DSP+FPGA frameworks are used, infrared image undercarriage automatic identification technology is realized in DSP Processor, meet the need handled in real time Ask, as shown in figure 1, specific implementation step is as follows：

(1) input infrared image I is obtained；

DSP Processor input picture I is 8 bit digital images, and picture size is 640 × 512.

(2) I histogram His are counted, the segmentation threshold Th of foreground and background is calculated；

Histogram His is expressed as His (k) { k=0...255 }, setting thresholding T₁=120, draw 1 × N of window and take N=10, root According to Count/N>0.8 can be calculated threshold value Th=130.

(3) image I binaryzations are obtained into binary map Bi Img, as shown in Fig. 2 b and Fig. 3 b using threshold value Th=130；

(4) positioning aircraft engine position Pos₁And Pos₂, i.e. (x₁,y₁) and (x₂,y₂)；

, will pixel count His (k) first in histogram His (k) { k=0...255 }>40 gray level is considered to start Machine gray value, can be calculated engine gray level for 240, according to gray value 240 can in infrared image location engine position Put Pos₁And Pos₂, for example in fig. 2 a be (320,304) and (398,296), in fig. 3 a for (158,412) and (182, 396)。

(5) undercarriage distributed areas are respectively Area before and after calculating₁、Area₂And Area₃；

Area₁It is represented by：Original position in x directions isEnd position in x directions isOriginal position in y directions isEnd position in y directions is

Area₂It is represented by：Original position in x directions is x₁；End position in x directions isIn y The original position in direction isEnd position in y directions is

Area₃It is represented by：Original position in x directions isEnd position in x directions is x₂；In y The original position in direction isEnd position in y directions is

Take length-width ratio T₂=3, according to above-mentioned expression formula, it can be calculated：

In fig. 2 a, Area₁Original position in x directions is 351；End position in x directions is 367；In y directions Original position is 300；End position in y directions is 326；Area₂It is represented by：Original position in x directions is 320；In x The end position in direction is 340；Original position in y directions is 300；End position in y directions is 326；Area₃It can represent For：Original position in x directions is 378；End position in x directions is 398；Original position in y directions is 300；In y side To end position be 326.

In fig. 3 a, Area₁Original position in x directions is 208；End position in x directions is 232；In y directions Original position is 404；End position in y directions is 445；Area₂It is represented by：Original position in x directions is 158；In x The end position in direction is 189；Original position in y directions is 404；End position in y directions is 445；Area₃It can represent For：Original position in x directions is 282；End position in x directions is 251；Original position in y directions is 404；In y side To end position be 445.

(6) in Area₁、Area₂And Area₃Calculate undercarriage characteristic parameter F₁、F₂And F₃；

According to F₁、F₂And F₃Computing formula can be obtained, in fig. 2 a, F₁=0.53, F₂=0.81 and F₃=0.79；In Fig. 3 a In, F₁=0.57, F₂=0.83 and F₃=0.85.

(7) according to characteristic parameter F₁、F₂And F₃Parameter completes undercarriage identification；

Take T₃=0.5, principle is recognized according to undercarriage, Fig. 2 a and Fig. 3 a undercarriage can be identified, such as Fig. 2 c and figure Shown in 3c.

The invention provides a kind of infrared image undercarriage automatic identifying method, implement the technical scheme method and Approach is a lot, and the above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should It is considered as protection scope of the present invention.Each part being not known in the present embodiment can use prior art to be realized.

Claims (6)

1. a kind of infrared image undercarriage automatic identifying method, it is characterised in that comprise the following steps：

Step 1, input infrared image I；

Step 2, statistics infrared image I histogram His, calculates the segmentation threshold Th of foreground and background；

Step 3, infrared image I binaryzations are obtained into binary map BiImg using segmentation threshold Th；

Step 4, the position Pos of two engines of aircraft is positioned₁And Pos₂；

Step 5, undercarriage distributed areas before and after calculating, respectively Area₁、Area₂And Area₃；

Step 6, fall into a trap calculation undercarriage characteristic parameter F in distributed areas₁、F₂And F₃；

Step 7, according to characteristic parameter F₁、F₂And F₃Complete undercarriage identification.

2. technology according to claim 1, it is characterised in that step 2 comprises the following steps：

Step 2-1, calculates infrared image I histogram His (k) { k=0 ..., 255 }, k represents gray level；

Step 2-2, setting thresholding T₁, 1 × N is carried out from high grade grey level to low gray level to histogram His and draws window processing, if drawing window The pixel count His (k) of interior current gray level>T₁, current gray level level is judged as background value, otherwise, it is determined that current gray level level is prospect Value, wherein, 5≤N≤15, N represents window width, 100≤T₁≤150；

Step 2-3,1 × N of statistics draws His (k) in window>T₁Number Count, if Count/N>0.8, then the k values at N/2 are sentenced It is set to segmentation threshold Th, otherwise, continues to carry out 1 × N strokes of window processing from high grade grey level to low gray level, until finding segmentation threshold Untill.

3. method according to claim 2, it is characterised in that step 4 includes：

Searched in histogram His from high grade grey level to low gray level, will pixel count His (k) first>40 gray level is determined as Engine gray value, the position Pos of two engines is positioned according to the gray value in infrared image₁And Pos₂, their coordinate Respectively (x₁,y₁) and (x₂,y₂), x₁,y₁Engine location Pos is represented respectively₁Abscissa and ordinate, x₂,y₂Represent respectively Engine location Pos₂Abscissa and ordinate.

4. method according to claim 3, it is characterised in that step 5 includes：

Undercarriage distribution characteristics is：Nose-gear region Area₁Near the axis of two engines, rear undercarriage region Area₂And Area₃It is located at the position Pos of two engines respectively₁And Pos₂Near, with two engine location coordinate (x₁,y₁) (x₂,y₂) it is reference point, x₁<x₂, rectangular coordinate system is set up, then：

Area₁It is expressed as：Original position in transverse axis x directions isEnd position in transverse axis x directions isIt is in the original position of longitudinal y directionsIt is in the end position of longitudinal y directions

Area₂It is expressed as：Original position in transverse axis x directions is x₁；End position in transverse axis x directions is The original position of longitudinal y directions isIt is in the end position of longitudinal y directions

Area₃It is expressed as：Original position in transverse axis x directions isEnd position in transverse axis x directions is x₂； The original position of longitudinal y directions isIt is in the end position of longitudinal y directions

Wherein, T₂Represent length-width ratio, 2≤T₂≤4。

5. technology according to claim 4, it is characterised in that step 6 includes：

In binary map BiImg, it is assumed that front and rear undercarriage distributed areas Area₁、Area₂And Area₃The y directions that pixel value is 1 are most Big position is respectively y_max1、y_max2And y_max3, then characteristic parameter F₁、F₂And F₃Expression formula is as follows：

$F_1=\left(\left|y_{max1}\right|-\frac{(y_1+y_2)}{2}\right)/\left(\frac{|x_1-x_2|}{T_2}\right);$

$F_2=\left(\left|y_{max2}\right|-\frac{(y_1+y_2)}{2}\right)/\left(\frac{|x_1-x_2|}{T_2}\right);$

$F_3=\left(\left|y_{max3}\right|-\frac{(y_1+y_2)}{2}\right)/\left(\frac{|x_1-x_2|}{T_2}\right).$

6. method according to claim 5, it is characterised in that step 7 includes：

If F₁>T₃, then judge that nose-gear is in down state, otherwise, it is determined that nose-gear is in collapsed state；

If F₂>T₃, then judge that rear left undercarriage is in down state, otherwise, it is determined that rear left undercarriage is in collapsed state；

If F₃>T₃, then right landing gear is in down state after judging, otherwise, it is determined that rear right landing gear is in collapsed state；

Wherein, 0.5≤T₃≤1。