A method for quickly identifying human targets by a vehicle-mounted infrared thermal imager

Abstract

The application discloses a kind of vehicle-mounted infrared thermal imager fast identification human target method, comprising: obtaining infrared thermogram in driving process, and it is carried out binaryzation processing;In the image after binaryzation processing, the target that meets template image is identified;Vehicle speed is read, the conversion relationship of the time required for the pixel height of vehicle speed and target in binaryzation image from a pixel height to the next pixel height changes is calculated;According to the conversion relationship, the interrelation table of the pixel height change and time change of target on infrared thermogram under different vehicle speed conditions is obtained;According to the interrelation table and the discrimination formula of human target when vehicle is driving, human target is judged;The application can effectively improve the identification rate and identification speed of vehicle-mounted infrared thermal imager to human target;Can identify long-distance human target, can guarantee driving safety.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of automobile auxiliary driving, and more particularly relates to a method for quickly identifying a human target by a vehicle-mounted infrared thermal imager. BACKGROUND

[0002] All objects in nature with a temperature higher than absolute zero (-273.15℃) radiate infrared rays at every moment, and the physical nature of infrared radiation is thermal radiation. Infrared thermal imaging is to discover and identify a target by using the infrared radiation characteristic image formed by the temperature difference or radiation difference between the target and the background or between different parts of the target. An infrared thermal imager is based on the principle of infrared thermal imaging, and focuses the infrared radiation of a scene in a field of view on an infrared detector through an infrared optical system capable of transmitting infrared radiation, so as to convert the infrared radiation energy into a physical quantity convenient for measurement. The infrared detector converts the radiation signals of different strengths into corresponding electrical signals, and then forms a video image for human eyes to observe on a display terminal after amplification and video processing.

[0003] The infrared thermal imager has the following advantages: 1. It is not affected by light changes and can solve the problem of dazzling of human eyes caused by strong light; 2. It is not affected by the strength of light and can solve the problem of limited vision at night; 3. It has strong penetration and can improve the problem of limited vision in adverse weather conditions such as rain, snow, fog, haze and dust. Due to its significant advantages, the infrared thermal imager has become an indispensable technology in the advanced driving assistance system (ADAS) of a vehicle, which can greatly improve the driving safety and comfort. At present, the vehicle-mounted infrared thermal imager mainly has two application scenarios: 1. To detect and display the video image of the field of view in front of the vehicle, and to enhance the driver's vision in adverse conditions; 2. To integrate algorithms to identify and calculate the distance of pedestrians in front of the vehicle, and to provide safety prompts to the driver when there is a potential collision danger.

[0004] At present, the traditional personnel identification algorithm based on the infrared thermal imager has achieved good results in some specific applications, but still has many deficiencies. For example, the human target cannot be effectively identified when the distance is far.

[0005] Most of these algorithms use deep learning methods, which directly update parameters by learning data, avoiding the design of complex algorithm processes by humans, and having high robustness and precision. However, the essence of deep learning is actually machine learning, which needs to provide enough data according to the principles of statistics, define the model, collect and analyze data, inject training to improve the model, output the result, and loop the above process to continuously improve the accuracy. In short, it is necessary to provide enough and various object surface templates for the algorithm to train and model. This puts higher requirements on the image processing capability of the infrared thermal imager in actual use, resulting in the increase of hardware cost and power consumption of the thermal imager.

[0006] The biggest disadvantage of the above method is that the required image resolution of the template is high, resulting in high requirements for computing resources. At the same time, under the condition that the focal length of the lens is unchanged, the high-resolution target image can only be collected at a relatively short distance, resulting in that the human target can only be identified at a relatively short distance.

[0007] Therefore, it is necessary to design a method for quickly identifying a human target of a vehicle-mounted infrared thermal imager. On the one hand, the human target can be quickly located and marked when the human target is far away from the vehicle, effectively improving the identification rate and speed of the vehicle-mounted infrared thermal imager for the human target. On the other hand, such an identification method should be simple and fast, only consuming a small amount of computing resources, effectively reducing the cost and power consumption of the equipment. SUMMARY

[0008] In view of the above defects or improvement needs of the prior art, the present application provides a method for quickly identifying a human target of a vehicle-mounted infrared thermal imager, which quickly identifies the human target according to the relationship between the vehicle speed and the time required for the pixel height of the target in the binary image to change from one pixel height to the next pixel height. The human target can be quickly located and marked when the human target is far away from the vehicle, effectively improving the identification rate and speed of the vehicle-mounted infrared thermal imager for the human target.

[0009] In order to achieve the above purpose, the present application provides a method for quickly identifying a human target of a vehicle-mounted infrared thermal imager, comprising the following steps:

[0010] S1: acquiring an infrared thermal image during driving by an infrared thermal imager, and performing binaryzation processing on the infrared thermal image to obtain a binary image;

[0011] S2: performing target identification in the binary image, and identifying the target in the image that meets the template image according to the pixel height of the target in the binary image;

[0012] S3: reading the vehicle speed, and calculating the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height;

[0013] S4: obtaining a correlation table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds according to the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height;

[0014] S5: performing human target determination according to the correlation table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds and the human target determination formula.

[0015] Further, the target recognition in the binary image in step S2 comprises the following steps:

[0016] selecting a typical image as a template image, and creating a template according to the gray value thereof;

[0017] measuring the relationship between the template image and the binary image in the processing module through a normalized cross-correlation matching algorithm, and recognizing and obtaining the target in the binary image that conforms to the template image.

[0018] Further, the reading of the vehicle speed in step S3 is realized through the processing module;

[0019] The obtaining of the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height in step S3 comprises the following steps:

[0020] S31: obtaining the distance between the infrared thermal imager and the measured target when the target presents N pixel point pixel heights on the infrared thermal image according to the lens focal length of the infrared thermal imager, the detector pixel spacing of the infrared thermal imager, and the height of the measured target;

[0021] S32: calculating the change of the distance between the vehicle and the measured target when the target presents N pixel point pixel heights to N+1 pixel point pixel heights on the infrared thermal image;

[0022] S33: calculating the time required for the target to present N pixel point pixel heights to N+1 pixel point pixel heights on the infrared thermal image according to the change of the distance between the vehicle and the measured target when the target presents N pixel point pixel heights to N+1 pixel point pixel heights on the infrared thermal image and the vehicle driving speed.

[0023] Further, the distance between the infrared thermal imager and the measured target when the target presents N pixel height on the infrared thermal image in step S31 is calculated by formula (4):

[0024]

[0025] Wherein, L is the distance between the infrared thermal imager and the measured target when the target presents N pixel height on the infrared thermal image; F is the focal length of the infrared thermal imager lens; H is the conventional target height; P is the pixel pitch of the infrared thermal imager detector; N is the pixel height of the target on the image display.

[0026] Further, the changed distance between the vehicle and the measured target when the target presents N pixel height to N+1 pixel height on the infrared thermal image in step S32 is calculated by formula (6):

[0027]

[0028] Wherein, ΔL is the changed value of the distance between the vehicle and the measured target when the target presents N pixel height to N+1 pixel height on the infrared thermal image.

[0029] Further, the time required for the target to present N pixel height to N+1 pixel height on the infrared thermal image in step S33 is calculated by formula (7):

[0030]

[0031] Wherein, Δt is the time required for the target to present N pixel height to N+1 pixel height on the infrared thermal image; V is the driving speed of the vehicle.

[0032] Further, step S5 further comprises the following steps:

[0033] S51: reading the driving speed of the vehicle, determining the starting pixel height of the human target according to the value formula of the starting pixel height of the human target under different driving speeds of the vehicle, and determining and marking the starting pixel height of the human target according to the correlation table of the pixel height change and the time change of the target on the infrared thermal image under different driving speeds of the vehicle.

[0034] S52: respectively calculating the time required for the pixel height of the target on the infrared thermal image to increase by one pixel point after the starting pixel height of the human target is determined, and comparing with the correlation table of the pixel height change and the time change of the target on the infrared thermal image under different driving speeds of the vehicle.

[0035] S53: if the comparison results of three times in succession all conform to the human target determination formula, it is determined that the current target is a human body.

[0036] Further, the value of the starting pixel height of the human target determined at different vehicle speeds in step S51 is represented by formula (8):

[0037]

[0038] where NV is the value of the starting pixel height of the target height determined at different vehicle speeds; and V is the vehicle speed.

[0039] Further, the time Δtl taken by the target to increase one pixel point in the pixel height on the infrared thermal image in step S52 is calculated by formula (9):

[0040]

[0041] The time Δt2 taken by the target to increase one pixel point in the pixel height on the infrared thermal image after increasing one pixel point in the pixel height is calculated by formula (10):

[0042]

[0043] The time Δt3 taken by the target to increase one pixel point in the pixel height on the infrared thermal image after increasing two pixel points in the pixel height is calculated by formula (11):

[0044]

[0045] Further, the human target determination formula in step S53 is represented by formula (12) to formula (14):

[0046]

[0047] where A is an array in the correlation table of the pixel height change and the time change of the target on the infrared thermal image at different vehicle speeds; and INT represents the function of the integer of the vehicle speed.

[0048] Q1 represents that the time taken by the target to increase one pixel point in the pixel height on the infrared thermal image after the starting pixel height of the human target is determined on the infrared thermal image is consistent with the numerical range of the correlation table of the pixel height change and the time change of the target on the infrared thermal image at different vehicle speeds.

[0049] Q2 represents that the time taken by the target to increase one pixel point in the pixel height on the infrared thermal image after increasing one pixel point in the pixel height is consistent with the numerical range of the correlation table of the pixel height change and the time change of the target on the infrared thermal image at different vehicle speeds.

[0050] Q3 represents: the time spent by the target on the infrared thermal image after the pixel height increases by two pixel points, and the numerical range of the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speed conditions.

[0051] Overall, compared with the prior art, the above technical solutions conceived by the present application can achieve the following beneficial effects:

[0052] 1. The method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to the present application quickly identifies a human target according to the relationship between the vehicle speed and the time required for the pixel height of the target in the binary image to change from one pixel height to the next pixel height, and can quickly locate and mark the human target when the human target is far away from the vehicle, thereby effectively improving the identification rate and speed of the vehicle-mounted infrared thermal imager for human targets.

[0053] 2. The method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to the present application is simple and fast, and only requires a small amount of computing resources, thereby effectively reducing the cost and power consumption of the equipment.

[0054] 3. The method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to the present application has a longer identification distance for human targets when the vehicle speed is higher, which meets the objective needs and can provide sufficient judgment time for the driver, thereby ensuring driving safety. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 Fig. 1 is a flowchart of the method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to an embodiment of the present application;

[0056] Figure 2 Fig. 2 is a flowchart of the acquisition of the conversion relationship between the vehicle speed and the time required for the pixel height of the target in the binary image to change from one pixel height to the next pixel height in the method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to an embodiment of the present application;

[0057] Figure 3 Fig. 3 is a flowchart of the determination of a human target according to the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speed conditions and the determination formula of the human target in the method for quickly identifying a human target of a vehicle-mounted infrared thermal imager according to an embodiment of the present application.

[0058] Abbreviations and key terms are defined as follows:

[0059] Field of view: referred to as the total field of view or the scanning field of view, indicating the maximum spatial angle range that can be observed when the infrared thermal imager is fixed in position.

[0060] Field of view angle algorithm:

[0061] The field of view angle is divided into horizontal field of view angle and vertical field of view angle:

[0062] Horizontal field of view angle=(pixel pitch / lens focal length)*horizontal pixel number / 17.45;

[0063] Vertical field of view angle=(pixel pitch / lens focal length)*vertical pixel number / 17.45.

[0064] Example:

[0065] Infrared thermal imager with a detector resolution of 640*480: 1) the detector pixel pitch used is 17μm; 2) a 40mm focal length lens is selected.

[0066] Therefore, the field of view angle of the infrared thermal imager with a detector resolution of 640*480 is:

[0067] Horizontal field of view angle:(17 / 40)*640 / 17.45=16°(generally rounded to the integer);

[0068] Vertical field of view angle:(17 / 40)*480 / 17.45=12°(generally rounded to the integer).

[0069] Spatial resolution definition: refers to the minimum distance of two adjacent targets that can be recognized by the infrared thermal imager, which is usually represented by the size of the instantaneous field of view angle (IFOV) (unit: milliradian mrad), indicating the minimum angular resolution unit of the infrared thermal imager.

[0070] Spatial resolution determines the definition of the infrared thermal imager, which is the smallest size that can be measured by the infrared thermal imager. It is related to the optical image quality, the focal length of the optical focusing system, and the linear size of the infrared sensor.

[0071] Spatial resolution algorithm: spatial resolution=pixel pitch / lens focal length; that is:

[0072] Where IFOV is the spatial resolution; P is the pixel pitch; F is the lens focal length;

[0073] Example:

[0074] The algorithm for the spatial resolution of the infrared thermal imager is:

[0075] 1) The detector pixel pitch used by the conventional infrared thermal imager is 17μm;

[0076] 2) The spatial resolution of the conventional infrared thermal imager with a 42mm focal length lens is: 17 / 42=0.405mrad;

[0077] 3) The spatial resolution of a conventional infrared thermal imager with a 25mm focal length lens is: 17 / 25 = 0.68 mrad. DETAILED DESCRIPTION

[0078] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application. In addition, the technical features involved in the various embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0079] The infrared thermal imager is based on the principle of infrared thermal imaging. Through an infrared optical system that can transmit infrared radiation, the infrared radiation of the scene in the field of view is focused on an infrared detector that can convert infrared radiation energy into a physical quantity that is easy to measure. The infrared detector converts the radiation signals of different strengths into corresponding electrical signals, and then amplifies and video processes them to form a video image that can be observed by the human eye on the display terminal. The infrared thermal imager is usually detachably fixed on a motorized carrier, and can real-time shoot the infrared thermal image of the motorized carrier driving road condition and output the video information of the infrared thermal image. However, most of the traditional personnel recognition algorithms based on the infrared thermal imager use deep learning methods, which need to provide enough and various object surface templates for algorithm training and modeling. This puts higher requirements on the image processing capability of the infrared thermal imager in actual use, resulting in the increase of hardware cost and power consumption of the thermal imager. At the same time, under the condition that the focal length of the lens is unchanged, the high-resolution target image can only be collected at a relatively short distance, resulting in that the human body target can only be identified at a relatively short distance. In the case that the human body target is at a relatively long distance, it cannot be effectively identified.

[0080] Based on the above reasons, as shown in Figure 1 The embodiment of the present application provides a vehicle-mounted infrared thermal imager rapid human body target recognition method, which is applied to a vehicle loaded with an infrared thermal imager to rapidly recognize a human body target during driving. The method comprises steps S1-S5.

[0081] S1: acquiring an infrared thermal image during driving by the infrared thermal imager, and performing binaryzation processing on the infrared thermal image to obtain a binaryzation image;

[0082] S2: performing target recognition in the binaryzation image, and identifying the target in the image that meets the template image according to the pixel height of the target presented in the binaryzation image;

[0083] S3: reading the vehicle speed, and calculating the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binaryzation image to change from one pixel height to the next pixel height;

[0084] S4: obtaining a correlation table of the pixel height change and time change of the target on the infrared thermal image under different vehicle speeds according to the correlation between the vehicle speed and the time required for the target to change from one pixel height to the next pixel height in the binary image;

[0085] S5: performing human target determination according to the correlation table of the pixel height change and time change of the target on the infrared thermal image under different vehicle speeds and the human target determination formula.

[0086] Further, in the embodiment of the present application, the infrared thermal image during driving is acquired by the infrared thermal imager in step 1, and the infrared thermal image is binary processed to obtain a binary image, and the embodiment further comprises the following steps: the infrared thermal image is binary processed by the processing module connected to the infrared thermal imager to obtain the binary image of the infrared thermal image.

[0087] Further, in the embodiment of the present application, the binary processing of the infrared thermal image by the processing module in step S1 comprises:

[0088] The processing module collects the ambient temperature data TAMB and inputs it into the infrared thermal imager;

[0089] The infrared thermal imager collects a frame of infrared thermal image data, and the processing module performs histogram statistics on the infrared thermal image data to obtain the effective lower limit level (GMIN) and upper limit level (GMAX) of the infrared thermal image data;

[0090] The processing module calculates the highest temperature (TMAX) and the lowest temperature (TMIN) in the infrared thermal image data;

[0091] The processing module converts the infrared thermal image data to generate a 256-level grayscale image one by one using the lower limit level (GMIN) and the upper limit level (GMAX) of the infrared thermal image data;

[0092] The processing module calculates the grayscale value corresponding to the ambient temperature;

[0093] The processing module performs binary processing on the grayscale image using the grayscale value corresponding to the ambient temperature;

[0094] The conversion of the infrared thermal image data to generate a 256-level grayscale image is represented by formula (1):

[0095] y = (BYTE) (((x-GMIN)*1.0)*255.0 / (GMAX-GMIN)) (1)

[0096] Wherein: x is the level data of infrared thermal image data; y is the converted gray data; (BYTE) indicates converting the floating point number obtained by formula operation into 8-bit integer (BYTE).

[0097] The processing module calculates the gray value GAMB corresponding to the ambient temperature, which is represented by formula (2):

[0098] GAMB=(TMAX-TAMB)*255 / (TMAX-TMIN) (2)

[0099] The processing module performs binaryzation processing on the gray image by using the gray value corresponding to the ambient temperature, which is represented by formula (3):

[0100]

[0101] Wherein, GYn represents the binaryzation data after the infrared thermal image data is processed by point-by-point binaryzation, and n represents the pixel point order number; the value range of n is 0-(image height*image width-1); Gn represents the gray value of the infrared image point by point, which is the same as the y value in formula (1).

[0102] Further, in the embodiment of the present application, the target recognition in the binaryzation image in step S2 includes the following steps:

[0103] A typical image is selected as a template image, and a template is created according to the gray value thereof;

[0104] The relationship between the template image and the binaryzation image in the processing module is measured by a normalized cross correlation matching algorithm, and the target in the binaryzation image that meets the template image is recognized and obtained;

[0105] Wherein, the normalized cross correlation matching (NCC) algorithm is used to measure the relationship between the template image and the detection image; the advantage is that it can retrieve some detection images with slight shape changes, complex textures or focus blur;

[0106] The relationship between the template image and the detection image is measured by the normalized cross correlation matching (NCC) algorithm, and the principle is that all the pixels in the template image are arranged in column order to form a row vector a, that is, the feature vector of the template, and then the most matching area b of the template is searched on the detection image, and the matching probability is measured by calculating the included angle θ between the two vectors, as shown below:

[0107]

[0108] The selection of the template image in the processing module meets the following condition: the target with a height-width ratio of 1:3.4 (50:1700) is identified in the image frame.

[0109] Further, in the embodiment of the present application, the reading of the vehicle speed in step S3 is realized by the processing module; as shown in the figure, the obtaining of the conversion relationship of the time required for the vehicle speed and the pixel height of the target presented in the binary image from one pixel height to the next pixel height comprises the following steps: Figure 2

[0110] S31: According to the focal length of the lens of the infrared thermal imager, the pixel pitch of the detector of the infrared thermal imager, and the height of the measured target, the distance between the infrared thermal imager and the measured target when the target presents N pixel heights on the infrared thermal image is obtained.

[0111] S32: The change of the distance between the vehicle and the measured target when the target presents N pixel heights to N+1 pixel heights on the infrared thermal image is calculated.

[0112] S33: According to the change of the distance between the vehicle and the measured target when the target presents N pixel heights to N+1 pixel heights on the infrared thermal image and the driving speed of the vehicle, the time required for the target to present N pixel heights to N+1 pixel heights on the infrared thermal image is calculated; and further, the conversion relationship of the time required for the vehicle speed and the pixel height of the target presented in the binary image from one pixel height to the next pixel height is obtained.

[0113] In step S31, the distance between the infrared thermal imager and the measured target when the target presents N pixel heights on the infrared thermal image is calculated by formula (5):

[0114]

[0115] Wherein, L is the distance between the infrared thermal imager and the measured target when the target presents N pixel heights on the infrared thermal image, with the unit of kilometers (km); F is the focal length of the lens of the infrared thermal imager, with the unit of millimeters (mm); H is the height of the conventional target, with the unit of meters (m); P is the pixel pitch of the detector of the infrared thermal imager, with the unit of micrometers (μm); N is the pixel height of the target displayed on the image, with the unit of pixel points.

[0116] In step S32, the change of the distance between the vehicle and the measured target when the target presents N pixel heights to N+1 pixel heights on the infrared thermal image is calculated by formula (6):

[0117]

[0118] ​Wherein, AL is the change value of the distance between the vehicle and the target on the infrared thermal image when the target presents the change from the pixel height of N pixel points to the pixel height of N+1 pixel points, and the unit is kilometer (km);

[0119] In step S33, the time required for the target to present the change from the pixel height of N pixel points to the pixel height of N+1 pixel points on the infrared thermal image is calculated by formula (7):

[0120]

[0121] Wherein, At is the time required for the target to present the change from the pixel height of N pixel points to the pixel height of N+1 pixel points on the infrared thermal image, and the unit is second (s); V is the driving speed of the vehicle, and the unit is kilometer per hour (km / h);

[0122] Therefore, the conversion relationship between the vehicle speed and the time required for the target to present the change from one pixel height to the next pixel height in the binary image is also represented by formula (7).

[0123] Further, in the embodiment of the present application, the conversion relationship between the vehicle speed and the time required for the target to present the change from one pixel height to the next pixel height in the binary image is obtained by the following examples.

[0124] Suppose:

[0125] The focal length of the infrared thermal imager lens: 18 mm;

[0126] The pixel pitch of the infrared thermal imager detector: 17 μm;

[0127] The resolution of the infrared thermal imager detector: 640×480;

[0128] The measured target: height 1.7 m;

[0129] The speed of the vehicle: V=60 km / h;

[0130] Then:

[0131] When the target presents 3 pixel points on the infrared thermal image:

[0132] The distance between the measuring instrument and the measured object is 1.7 / ((17 / 18)×3)=0.600 km;

[0133] When the human body presents 4 pixel points in the image:

[0134] The distance between the measuring instrument and the measured object is 1.7 / ((17 / 18)×4)=0.450 km;

[0135] When the human body presents 5 pixel points in the image:

[0136] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 5) = 0.360 km;

[0137] When the human body presents 6 pixel points in the image:

[0138] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 6) = 0.300 km;

[0139] When the human body presents 7 pixel points in the image:

[0140] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 7) = 0.257 km;

[0141] When the human body presents 8 pixel points in the image:

[0142] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 8) = 0.225 km;

[0143] When the human body presents 9 pixel points in the image:

[0144] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 9) = 0.200 km;

[0145] When the human body presents 10 pixel points in the image:

[0146] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 10) = 0.180 km;

[0147] When the human body presents 11 pixel points in the image:

[0148] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 11) = 0.164 km;

[0149] When the human body presents 12 pixel points in the image:

[0150] The distance between the measuring instrument and the measured object = 1.7 / ((17 / 18) x 12) = 0.150 km;

[0151] Taking the change of the distance ΔS between the vehicle and the measured target from 3 pixel points to 4 pixel points as an example, the calculation is as follows:

[0152] ΔS = 1.7 / ((17 / 18) x 3) - 1.7 / ((17 / 18) x 4) = 1.7 / (17 / 18) / ((4 x 3) / (4-3));

[0153] It is also known that ΔS = V x T, T is the time required for the human body to change from 3 pixel points to 4 pixel points on the image, with a unit of seconds (s); the unit of ΔS is km;

[0154] According to 3600 seconds for 1 hour, it can be known that:

[0155]

[0156] Further, reverse verification:

[0157] If the vehicle speed V = 60 km / h at this time, then:

[0158] From 3 pixel points to 4 pixel points, the vehicle is close to the measured target by 0.15 km, according to the distance-speed-time formula, the time t = 9s can be calculated;

[0159] From 4 pixel points to 5 pixel points, the vehicle is close to the measured target by 0.09 km, according to the distance-speed-time formula, the time t = 5.4s can be calculated;

[0160] From 5 pixel points to 6 pixel points, the vehicle is close to the measured target by 0.06 km, according to the distance-speed-time formula, the time t = 3.6s can be calculated;

[0161] From 6 pixel points to 7 pixel points, the vehicle is close to the measured target by 0.043 km, according to the distance-speed-time formula, the time t = 2.5s can be calculated;

[0162] From 7 pixel points to 8 pixel points, the vehicle is close to the measured target by 0.032 km, according to the distance-speed-time formula, the time t = 1.8s can be calculated;

[0163] From 8 pixel points to 9 pixel points, the vehicle is close to the measured target by 0.025 km, according to the distance-speed-time formula, the time t = 1.44s can be calculated;

[0164] From 9 pixel points to 10 pixel points, the vehicle is close to the measured target by 0.02 km, according to the distance-speed-time formula, the time t = 1.08s can be calculated, and the image changes 25 x 1.08 = 27 frames;

[0165] From 10 pixel points to 11 pixel points, the vehicle is close to the measured target by 0.016 km, according to the distance-speed-time formula, the time t = 0.972s can be calculated;

[0166] From 11 pixel points to 12 pixel points, the vehicle is close to the measured target by 0.014 km, according to the distance-speed-time formula, the time t = 0.828s can be calculated;

[0167] Therefore, according to the above calculation process, a conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height can be obtained.

[0168] Further, in the embodiment of the present application, in step S4, the mutual relationship table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds is shown in Table 1.

[0169] Table 1 Mutual relationship table of pixel height change and time change of target on infrared thermal image

[0170] 11-12 0 0 0 1.69 1.26 1.01 0.83 0.72 0.63 0.56 0.50 10-11 0 0 0 1.91 1.44 1.15 0.97 0.82 0.72 0.64 0.58 9-10 0 0 0 2.41 1.80 1.44 1.08 1.03 0.90 0.80 0.72 8-9 0 0 0 2.99 2.25 1.80 1.44 1.29 1.13 1.00 0.90 7-8 0 0 0 3.85 2.88 2.30 1.80 1.65 1.44 1.28 1.15 6-7 0 0 0 5.15 3.87 3.10 2.50 2.21 1.94 1.72 1.55 5-6 0 0 0 7.20 5.40 4.32 3.60 3.09 2.70 2.40 2.16 4-5 0 0 0 10.80 8.10 6.48 5.40 4.63 4.05 3.60 3.24 3-4 0 0 0 18.00 13.50 10.68 9.00 7.71 6.75 6.00 5.40 2-3 0 0 0 0 0 0 0 0 0 0 0 1-2 0 0 0 0 0 0 0 0 0 0 0 0-1 0 0 0 0 0 0 0 0 0 0 0 0 10 20 30 40 50 60 70 80 90 100 Speed

[0171] A two-dimensional array A (11, 12) is constructed according to Table 1, wherein the data in the X-axis direction represents the speed V of the vehicle driving, and the Y-axis direction represents the pixel height change of the target height displayed on the infrared thermal imager.

[0172] Further, in the embodiment of the present application, as shown in Table 1, the human target determination in step S5 is performed according to the mutual relationship table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds and the human target determination formula when the vehicle is driving, and includes the following steps: Figure 3

[0173] S51: reading the vehicle driving speed, according to the value formula of the starting pixel height of the human target determination under different vehicle driving speeds, comparing with the mutual relationship table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds, determining the starting pixel height of the human target determination and marking;

[0174] S52: respectively calculating the time required for the pixel height of the human target on the infrared thermal image to increase by one pixel point after the starting pixel height of the human target determination, and comparing with the mutual relationship table of the pixel height change and the time change of the target under different vehicle speeds;

[0175] S53: if the comparison results of three times in succession all conform to the human target determination formula, it is determined that the current target is a human body.

[0176] Further, in the embodiment of the present application, the value of the starting pixel height of the human target determination under different vehicle driving speeds in step S51 is represented by formula (8):

[0177]

[0178] ​Wherein, NV is the value of the starting pixel height of the starting height of the target at different vehicle speeds; V is the vehicle speed; that is:

[0179] When the read vehicle speed is 30≤V<40, the value of the starting pixel height of the starting height of the target is 10 pixel points;

[0180] When the read vehicle speed is 40≤V<50, the value of the starting pixel height of the starting height of the target is 9 pixel points;

[0181] When the read vehicle speed is 50≤V<60, the value of the starting pixel height of the starting height of the target is 8 pixel points;

[0182] When the read vehicle speed is 60≤V<70, the value of the starting pixel height of the starting height of the target is 7 pixel points;

[0183] When the read vehicle speed is 70≤V<80, the value of the starting pixel height of the starting height of the target is 6 pixel points;

[0184] When the read vehicle speed is 80≤V<90, the value of the starting pixel height of the starting height of the target is 5 pixel points;

[0185] When the read vehicle speed is 90≤V<100, the value of the starting pixel height of the starting height of the target is 4 pixel points.

[0186] Further, in the embodiment of the present application, the pixel point corresponding to the starting pixel height of the human target is defined as the initial determination pixel height; the pixel point after the starting pixel height of the human target is increased by one pixel point is recorded as the first pixel height; the pixel point after the starting pixel height of the human target is increased by two pixel points is recorded as the second pixel height; the pixel point after the starting pixel height of the human target is increased by three pixel points is recorded as the third pixel height; then:

[0187] The time spent by the target on the infrared thermal image to increase the pixel height by one pixel point is the time spent by the target from the initial determination pixel height to the first pixel height;

[0188] The time spent by the target on the infrared thermal image to increase the pixel height by one pixel point after increasing the pixel height by one pixel point is the time spent by the target from the first pixel height to the second pixel height;

[0189] The time spent by the target on the infrared thermal image to increase the pixel height by one pixel point after increasing the pixel height by two pixel points is the time spent by the human target from the second pixel height to the third pixel height;

[0190] The time Δti spent by the target in increasing the pixel height on the infrared thermal image by one pixel point (i.e. the time spent by the target in increasing the pixel height from the initial determined pixel height to the first pixel height) in step S52 is calculated by formula (9):

[0191]

[0192] The time Δt2 spent by the target in increasing the pixel height on the infrared thermal image by one pixel point after increasing the pixel height by one pixel point (i.e. the time spent by the target in increasing the pixel height from the first pixel height to the second pixel height) is calculated by formula (10):

[0193]

[0194] The time Δt3 spent by the target in increasing the pixel height on the infrared thermal image by one pixel point after increasing the pixel height by two pixel points (i.e. the time spent by the target in increasing the pixel height from the second pixel height to the third pixel height) is calculated by formula (11):

[0195]

[0196] Further, in the embodiment of the present application, the human target determination formula in step S53 is represented by formula (12) to formula (14):

[0197]

[0198] Wherein, A is an array in the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds; INT represents the function of rounding the vehicle speed;

[0199] Q1 represents that the time spent by the target in increasing the pixel height on the infrared thermal image by one pixel point after determining the starting pixel height of the human target on the infrared thermal image is consistent with the numerical range of the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds;

[0200] Q2 represents that the time spent by the target in increasing the pixel height on the infrared thermal image by one pixel point after increasing the pixel height by one pixel point is consistent with the numerical range of the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds;

[0201] Q3 represents that the time spent by the target in increasing the pixel height on the infrared thermal image by one pixel point after increasing the pixel height by two pixel points is consistent with the numerical range of the table of the mutual relationship between the pixel height change and the time change of the target on the infrared thermal image under different vehicle speeds;

[0202] Q1, Q2, Q3 are all logical values determined by the starting pixel height of the human target meeting the value formula condition of the starting pixel height of the human target.

[0203] From the above, it can be seen that the height of the selected starting pixel point is not the same at different vehicle speeds, and the selection principle is to ensure that the time interval of the pixel height increase of the target on the infrared thermal image is controlled within 2 seconds. According to this principle, when the vehicle speed is 30 km / h, the starting judgment value of the height of the target is 10 pixels; when the vehicle speed is 100 km / h, the starting judgment value of the height of the target is 5 pixels; according to Table 1 and the relationship between the pixel and the distance, it can be concluded that when the vehicle speed is low (for example, 30 km / h), the vehicle-mounted infrared thermal imager fast recognition human target method of the present application can start human body discrimination from 180 meters away; when the vehicle speed is high (for example, 100 km / h), the vehicle-mounted infrared thermal imager fast recognition human target method of the present application can start human body discrimination from 360 meters away; it can provide sufficient judgment time for the driver; it can be seen that the vehicle-mounted infrared thermal imager fast recognition human target method of the present application can recognize the human target at a farther distance when the vehicle speed is higher, which can fully guarantee the safety of driving.

[0204] Those skilled in the art can easily understand that the above description is only a preferred embodiment of the present application and is not intended to limit the present application, and any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for quickly identifying human targets using a vehicle-mounted infrared thermal imager, characterized in that: The steps include: S1: Acquire an infrared thermal image during driving by using an infrared thermal imager, and perform binarization processing on the infrared thermal image to obtain a binary image; S2: Target recognition is performed in the binary image. Targets that match the template image are identified based on the pixel height of the target in the binary image. S3: Read the vehicle speed and calculate the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height; In step S3, obtaining the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height includes the following steps: S31: according to the lens focal length of the infrared thermal imager, the pixel pitch of the infrared thermal imager detector and the height of the target to be measured, obtaining the distance between the infrared thermal imager and the target to be measured when the target appears at a pixel height of N pixels on the infrared thermal image; S32: Calculating the change in the distance between the vehicle and the target when the target changes from a pixel height of N pixels to a pixel height of N+1 pixels on the infrared thermal image; S33: Calculating the time required for the target to change from N pixels to N+1 pixels in height on the infrared thermal image based on the change in the distance between the vehicle and the target and the vehicle's speed when the target changes from N pixels to N+1 pixels in height on the infrared thermal image; S4: Based on the conversion relationship between the vehicle speed and the time required for the pixel height of the target presented in the binary image to change from one pixel height to the next pixel height, a correlation table of the pixel height change and the time change of the target on the infrared thermal image under different vehicle speed conditions is obtained; S5: Human target determination is performed based on a correlation table of pixel height changes and time changes of targets on infrared thermal images at different vehicle speeds and a human target determination formula when the vehicle is moving.

2. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 1, characterized in that: In step S2, target recognition in the binary image includes the following steps: Select a typical image as a template image and create a template based on its grayscale value; The normalized cross-correlation matching algorithm is used to measure the relationship between the template image and the binary image in the processing module, and the target in the binary image that matches the template image is identified and obtained.

3. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 2, characterized in that: Reading the vehicle speed in step S3 is achieved through the processing module.

4. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 3, characterized in that: In step S31, the distance between the infrared thermal imager and the target when the target appears at N pixel heights on the infrared thermal image is calculated using formula (5): Where L is the distance between the infrared thermal imager and the target when the target appears N pixels high on the infrared thermal image; F is the focal length of the infrared thermal imager lens; H is the normal target height; P is the pixel pitch of the infrared thermal imager detector; and N is the pixel height of the target displayed on the image.

5. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 4, characterized in that: In step S32, the distance between the vehicle and the target when the target changes from N pixels to N+1 pixels on the infrared thermal image is calculated by formula (6): ΔL is the change in the distance between the vehicle and the target when the target changes from N pixels to N+1 pixels on the infrared thermal image.

6. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 5, characterized in that: In step S33, the time required for the target to appear on the infrared thermal image from the pixel height of N pixels to the pixel height of N+1 pixels is calculated by formula (7): Wherein, Δt is the time required for the target to appear from N pixels to N+1 pixels in height on the infrared thermal image; V is the speed of the vehicle.

7. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 6, characterized in that: Step S5 also includes the following steps: S51: Read the vehicle speed, determine the starting pixel height for determining a human target based on a formula for determining the starting pixel height for different vehicle speeds, and compare the relationship between the pixel height change and time change of the target on the infrared thermal image at different vehicle speeds to mark the starting pixel height for determining a human target; S52: Calculating the time required for the pixel height of the target to increase by one pixel after the starting pixel height for determining a human target on the infrared thermal image, and comparing the calculated time with the correlation table of pixel height change and time change of the target on the infrared thermal image under different vehicle speed conditions; S53: If the comparison results for three consecutive times all meet the human target determination formula, the current target is determined to be a human body.

8. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 7, characterized in that: The values ​​of the starting pixel height for determining the human target at different vehicle speeds in step S51 are expressed by formula (8): Among them, NV is the value of the starting pixel height for determining the target height at different vehicle speeds; V is the vehicle speed.

9. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 8, characterized in that: In step S52, the time Δt1 taken for the pixel height of the target on the infrared thermal image to increase by one pixel is calculated by formula (9): The time Δt2 required for the target to increase its pixel height by one pixel on the infrared thermal image is calculated by formula (10): The time Δt3 required for the target to increase its pixel height by two pixels and then increase by one pixel on the infrared thermal image is calculated by formula (11):

10. The method for quickly identifying human targets using a vehicle-mounted infrared thermal imager according to claim 9, characterized in that: The human target determination formula in step S53 is expressed by equations (12) to (14); Where A is an array in the table of the relationship between the pixel height change and time change of the target on the infrared thermal image under different vehicle speeds; INT represents the function of rounding the vehicle speed; Q1 indicates that the time it takes for the pixel height to increase by one pixel after the target determines the starting pixel height of the human target on the infrared thermal image is consistent with the numerical range of the correlation table of the target pixel height change and time change on the infrared thermal image under different vehicle speeds; Q2 indicates that after the pixel height of the target on the infrared thermal image increases by one pixel, the time it takes to increase by one more pixel is consistent with the numerical range of the correlation table of the pixel height change and time change of the target on the infrared thermal image under different vehicle speeds; Q3 indicates that after the pixel height of the target on the infrared thermal image increases by two pixels, the time it takes to increase one pixel is consistent with the numerical range of the correlation table between the pixel height change and time change of the target on the infrared thermal image under different vehicle speeds.

Images