WO2018120443A1 - 一种点斑状目标姿态估计方法及系统 - Google Patents

一种点斑状目标姿态估计方法及系统 Download PDF

Info

Publication number
WO2018120443A1
WO2018120443A1 PCT/CN2017/077101 CN2017077101W WO2018120443A1 WO 2018120443 A1 WO2018120443 A1 WO 2018120443A1 CN 2017077101 W CN2017077101 W CN 2017077101W WO 2018120443 A1 WO2018120443 A1 WO 2018120443A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
spot
radiant energy
infrared
radiation
Prior art date
Application number
PCT/CN2017/077101
Other languages
English (en)
French (fr)
Inventor
张天序
王凤林
姚守悝
陆檑
程旭
Original Assignee
华中科技大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华中科技大学 filed Critical 华中科技大学
Publication of WO2018120443A1 publication Critical patent/WO2018120443A1/zh
Priority to US16/458,221 priority Critical patent/US11162847B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J5/0022—Radiation pyrometry, e.g. infrared or optical thermometry for sensing the radiation of moving bodies
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J5/02—Constructional details
    • G01J5/07—Arrangements for adjusting the solid angle of collected radiation, e.g. adjusting or orienting field of view, tracking position or encoding angular position
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J2005/0077—Imaging
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J5/48—Thermography; Techniques using wholly visual means

Definitions

  • the invention belongs to the technical field of target attitude estimation, and particularly relates to a point spot target attitude estimation method and system.
  • Attitude estimation of 3D targets is a hot research topic in the field of computer vision. It is an important part of 3D target motion estimation. It has been extensively studied in face recognition, aircraft operation and control. Moreover, the three-dimensional attitude is an important parameter reflecting the motion state of the space target, which is beneficial to the optimization design and fault analysis of the target, and also has important significance in the application of target tracking, target passive positioning and target recognition.
  • the target has a low spatial resolution and is imaged as a spotted target, at this time, the three-dimensional posture information of the target cannot be accurately known only according to the spotted target. Therefore, how to quickly and accurately estimate the posture of a spotted target has become an urgent problem to be solved.
  • attitude estimations for three-dimensional targets at home and abroad.
  • the existing common methods for measuring the attitude of a space target are mainly divided into two types: the inner side of the space target and the outer side of the space target. It is to track the video through the device, analyze the target information in the time series, and estimate the three-dimensional pose parameters. Obviously, this method is too costly. In addition, most of them are estimated for the surface targets that can reflect the target shape information, and the pose targets are not estimated. Moreover, the method of processing is directed to the physical information of the target itself, and the relationship of the infrared radiation characteristic curve of the target with the attitude change is not considered.
  • the present invention proposes a spotted target pose
  • the estimation method establishes a characteristic database corresponding to the spectrum and the attitude, analyzes the radiation characteristics of each component of the target in different postures, and then roughly estimates the posture of the spotted target according to the different forms of the measured infrared spectrum, and solves how to according to the infrared Spectral information estimates the problem of point-like target poses.
  • a point-like target attitude estimation method comprising an offline training part and an online estimation part
  • the offline training portion includes the following steps:
  • the infrared spectrum curve of the spot-like target radiant energy of the wavelength-target image is simulated and calculated, so as to establish a mapping database of target pose and spectrum;
  • the online estimation section includes the following steps:
  • step (4) is:
  • ⁇ is the wavelength
  • T (x, y, z) is the target object temperature at the position (x, y, z)
  • the black body emissivity ⁇ 1, The infrared emissivity of the target surface material
  • ⁇ ⁇ is the transmittance and L ⁇ is the atmospheric radiation.
  • the target is an aircraft
  • the radiance of the aircraft is mainly divided into three parts: an aircraft skin, an engine, and a tail flame.
  • the target object temperature T (x, y, z) at the position (x, y, z) is first processed to obtain the target object temperature after considering the speed:
  • k is the recovery coefficient
  • ⁇ is the ratio of the constant pressure heat capacity of the air to the constant volume heat capacity
  • M is the speed.
  • the specific implementation manner of the step (8) is: firstly determining the attitude of the aircraft according to the spot-like target infrared spectrum curve of the wavelength-target image radiant energy obtained in the step (7), and extracting the aircraft posture from the mapping database A set of spectral curves corresponding to the pose; then an exact match is made in the set of spectral curves to determine the target pose.
  • the target is an aircraft, and the infrared spectrum curve of the spotted target with respect to the wavelength-target image radiant energy obtained in the step (7) is analyzed:
  • the initial judgment posture is head-on
  • the initial determination posture is the side
  • the initial attitude is determined as a top view or a bottom view
  • the initial judgment is the tail chase.
  • a point-like target attitude estimation system comprising an offline training part and an online estimation part;
  • the offline training portion includes the following modules:
  • a three-dimensional model building module for establishing a target three-dimensional geometric model and dividing the region according to the target structure
  • An object temperature distribution model establishing module is configured to establish an object temperature distribution model for each region of the target
  • the infrared radiation transmission model establishing module is configured to establish a detection system to observe an infrared radiation transmission model of the target in the atmosphere in six typical postures; the six postures include top view, top view, head-on, rear-end, left and right, and side;
  • the image radiant energy model building module is used to construct the image radiant energy model of the target in six typical poses by using the established object temperature distribution model and the infrared radiation transmission model;
  • the mapping database building module is used to combine the target image radiant energy model in six typical poses, and simulate the calculated infrared spectrum curve of the spot-like target radiant energy of the wavelength-target image, thereby establishing the mapping of the target pose and the spectrum. database;
  • the online estimation section includes the following modules:
  • a detection module for detecting a spotted target
  • a spot-shaped target infrared spectrum curve generating module is used for real-time collecting spot-like target image radiant energy, and drawing a spot-like target infrared spectrum curve about wavelength-target image radiant energy;
  • the attitude estimation module is configured to match the generated spot-shaped target infrared spectrum curve about the wavelength-target image radiant energy in a mapping database of the target pose and the spectrum to determine the target pose.
  • the image radiant energy model building module includes:
  • the radiation characteristic expression of the target object constructs a sub-module, and the radiation characteristic expression L(T (x, y, z) ) of the target object is constructed according to the object temperature distribution model:
  • ⁇ is the wavelength
  • T (x, y, z) is the target object temperature at the position (x, y, z)
  • the black body emissivity ⁇ 1, The infrared emissivity of the target surface material
  • ⁇ ⁇ is the transmittance and L ⁇ is the atmospheric radiation.
  • attitude estimation module includes:
  • a preliminary determining sub-module configured to initially determine an aircraft attitude according to the generated spot-like target infrared spectrum curve of the wavelength-target image radiant energy, and extract a set of spectral curves corresponding to the posture from the mapping database;
  • the exact matching sub-module is used to accurately match the set of spectral curves to determine the target pose.
  • the target pose is effectively divided into six typical poses, which are simple, reasonable and representative.
  • the target is three-dimensional in spatial distribution, and can represent multiple posture information of the target under different observation angles.
  • the posture of the target is divided into six postures: head-on, tail-following, looking up, looking down, and left and right side views, and then a spectral database corresponding to the six postures is established.
  • the posture of the proposed target is related to the shape of the observed spectral curve.
  • the infrared radiation spectrum of the target is mainly determined by the temperature, and the material structure and composition of the target itself have an important influence on the temperature at each point on the target surface. That is, the temperature of the target can be divided into several different regions according to the structure, and the target regions observed by the target in different postures are also different, so that the spectral curves of the observed spotted targets are also different.
  • the present invention considers the temperature distribution and radiation characteristics of the target as a function of various positions such as spatial position, attitude, and detection distance; the target posture is effectively distinguished, and the distinction is simple and reasonable, and the accuracy is high. Solved the problem of how to estimate the attitude of a spotted target based on infrared spectral information;
  • 1 is a flow chart of a method for estimating a point target of a spotted target
  • Figure 2 is a schematic diagram of the spectral characteristics of the aircraft
  • Figure 3 is a graph showing the average spectral characteristics of the aircraft and the sky, the earth, and the sun;
  • Figure 4 is a schematic diagram of aircraft spectral data at different azimuth angles
  • Figure 5 is a spectrum diagram of SU-27 radiation in a certain attitude
  • Figure 6 is a schematic diagram showing the relative positional relationship of the intra-atmosphere detection system when observing targets in the atmosphere.
  • a point-like target pose estimation method includes an offline training part and an online evaluation part.
  • the offline training part includes establishing a three-dimensional geometric model, a temperature distribution model and an atmospheric transmission model of the target; setting simulation parameters, simulating the infrared radiation spectrum of the spotted target in six postures, and establishing a “attitude-spectrum” database;
  • the online evaluation part includes The measured infrared spectral morphology is analyzed and the current pose of the target is estimated by comparing the "attitude-spectrum” database.
  • FIG. 1 The process of the method of the present invention is shown in FIG. 1 , taking a certain type of aircraft as an example, and the specific implementation method includes the following steps:
  • the 3D model of the aircraft was built using 3dsmax software, and the different parts of the aircraft were correctly distinguished.
  • the aircraft parts mainly include the tail flame, the engine, the leading edge of the wing, the trailing edge of the wing, the nose, and the general part of the cockpit.
  • An object-side temperature distribution model is established for each region of the target, which can be measured in real time or simulated, or a fitting method can be employed.
  • the present invention preferably adopts an interpolation fitting method. Specifically, according to the obtained aircraft temperature data, the temperature data of the different regions are subjected to interpolation fitting processing, thereby approximating the temperature distribution function of the spatial position change in each region, and then The temperature of the target point at other positions in each region is calculated according to the function obtained by the fitting.
  • the detection system is respectively established to observe the infrared radiation transmission mode of the target in the atmosphere under six typical attitudes.
  • the establishment process is as follows:
  • the intracavity detection system observes the radiation transmission path under the target conditions in the atmosphere as shown in Fig. 5: the position of the detection system in the atmosphere is represented by point A, and the position of the target in the observed atmosphere is represented by point B, and the radiation of the observed object Energy travels from point B to point A.
  • the input parameters are: detection system height, target height, slant distance between the detection system and the target, the highest altitude of the atmosphere, and the zenith angle. , Earth radius, infrared band. The height of the detection system and the infrared band are determined according to the actual situation; the radius of the earth is the radius at the latitude of the detection system; the highest altitude of the atmosphere is set to 100 km; the zenith angle is shown by ⁇ CAB in Figure 6, and its calculation formula for:
  • the transmittance ⁇ ⁇ and the path radiation L ⁇ of the observation target of the detection system in the atmosphere at the angle ⁇ can be obtained.
  • the radiation characteristic expression L(T (x, y, z) ) of the target object is constructed according to the object temperature distribution model.
  • the infrared emissivity of the target surface material is the infrared emissivity of the target surface material.
  • the radiation characteristic expression L(T (x, y, z) , ⁇ ) constructed on the target image side is considered in consideration of the atmospheric influence.
  • the infrared radiation composition of the aircraft is as shown in FIGS. 2 and 3, and therefore the radiance of the aircraft skin, the engine, the tail flame, and the like is mainly calculated.
  • the specific calculation is as follows:
  • the surface temperature of the aircraft is related to the flight speed (Mach number M). First, T (x, y, z) should be considered in consideration of the speed factor, and then the skin is calculated using the formula L(T (x, y, z) , ⁇ ). Radiation intensity at the place.
  • M - speed is the flying Mach number.
  • the pixel gradation information is obtained from the spotted target image, and the corresponding image radiant energy is obtained by the calibration method, and the infrared spectroscopy curve of the punctured target is obtained according to the image radiant energy of the multi-point.
  • the aircraft attitude is initially determined according to the following typical attitude spectral characteristics, and the spectral curve set corresponding to the attitude is extracted from the database.
  • Head-on attitude Observed from the nose of the aircraft to the tail.
  • the spectrum of the point target formed at this time is dominated by the long-wave infrared radiation of the skin, and other components are less.
  • the spot spectral target infrared spectrum curve is accurately matched to determine the target pose.
  • the infrared radiation spectrum of the SU-27 aircraft actually measured is qualitatively analyzed and quantitatively calculated.
  • the spectrum has a radiation component in the short wave, medium wave and long wave, and the medium and long wave is in the range of 5.5 ⁇ m to 7.5 ⁇ m. There is no radiation, and the medium wave radiation is the strongest, the short wave is the second, and the long wave radiation is the weakest.
  • the SU-27 aircraft measured at this time is in the tail-following attitude.
  • the main observation is the tail flame and the tail nozzle, while the temperature of the tail flame and the tail nozzle is the surface of the aircraft. The higher the temperature of the point is also in line with the theoretical analysis.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Medical Informatics (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Radiation Pyrometers (AREA)
  • Image Processing (AREA)
  • Aiming, Guidance, Guns With A Light Source, Armor, Camouflage, And Targets (AREA)

Abstract

一种点斑状目标姿态估计方法及系统，属于目标姿态估计技术领域。该方法包括离线训练部分和在线评估部分；离线训练部分包括建立目标的三维几何模型、温度分布模型及大气传输模型；结合上述模型，仿真计算六种姿态下点斑状目标的红外辐射光谱，建立姿态-光谱映射数据库；在线评估部分测量红外光谱形态，比对姿态-光谱映射数据库估计该目标的当前姿态。该方法通过分析目标的光谱曲线并匹配光谱数据库反推目标当前所处的姿态，简单有效，解决了如何根据红外光谱信息估计点斑状目标姿态的难题。

Description

一种点斑状目标姿态估计方法及系统 [技术领域]
本发明属于目标姿态估计技术领域，具体涉及一种点斑状目标姿态估计方法及系统。
[背景技术]
三维目标的姿态估计是计算机视觉领域的一个热点研究课题，是三维目标运动估计的重要部分，在人脸识别、飞行器的操作与控制等方面都有广泛的研究。并且，三维姿态是反映空间目标运动状态的重要参数，有利于对目标的优化设计和故障分析，在目标跟踪、目标被动定位和目标识别等应用方面也具有重要的意义。
当目标在空间分辨率很低、成像为点斑状目标的情况下，此时仅根据点斑状目标无法准确知道目标的三维姿态信息。因此，如何进行快速准确地估计出点斑状目标的姿态已成为人们迫切需要解决的难题。
目前，在国内外也有一些针对三维目标的姿态估计。已有的测定空间目标姿态的常用方法主要分为空间目标姿态内侧和空间目标姿态外侧两种。它是通过设备进行跟踪录像，分析在时间序列上的目标信息，从而估计三维姿态参数，显然这种方式成本太高。另外大都是针对能够反映出目标形状信息的面目标进行估计，并未对点斑状目标进行姿态估计。而且所处理的方式均是针对目标本身物理信息，未考虑目标的红外辐射特性曲线随着姿态变化的关系。
[发明内容]
针对现有技术的迫切技术需求，本发明提出了一种点斑状目标姿态 估计方法，该方法建立光谱与姿态对应的特性数据库，分析目标在不同姿态下各组成部分的辐射特性，然后根据所测红外谱的形态不同从而粗略估计点斑状目标的姿态，解决了如何根据红外光谱信息估计点斑状目标姿态的难题。
一种点斑状目标姿态估计方法，包括离线训练部分和在线估计部分；
所述离线训练部分包括以下步骤：
(1)建立目标三维几何模型并依据目标结构划分区域；
(2)针对目标的各区域建立物方温度分布模型；
(3)建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模型；所述六种姿态包括顶视、俯视、迎头、追尾、左右、侧面；
(4)利用已建的物方温度分布模型和红外辐射传输模型构建目标在六种典型姿态下的像方辐射能量模型；
(5)分别在六种典型姿态下，结合目标像方辐射能量模型，仿真计算得到关于波长-目标像方辐射能量的点斑状目标红外光谱曲线，从而建立目标姿态与光谱的映射数据库；
所述在线估计部分包括以下步骤：
(6)探测点斑状目标；
(7)实时采集点斑状目标像方辐射能量，绘制关于波长-目标像方辐射能量的点斑状目标红外光谱曲线；
(8)将步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线在目标姿态与光谱的映射数据库内进行匹配，确定目标姿态。
进一步，所述步骤(4)的具体实现过程为：
(41)根据物方温度分布模型构建目标物方的辐射特性表达式L(T(x,y,z))：已知在波段λ1～λ2、温度为T的等效黑体辐射亮度Lb(T(x,y,x))为：
Figure PCTCN2017077101-appb-000001
则目标表面红外辐射亮度L(T(x,y,z))为：
Figure PCTCN2017077101-appb-000002
式中：λ为波长，T(x,y,z)为在位置(x,y,z)处的目标物方温度，黑体发射率ε＝1，
Figure PCTCN2017077101-appb-000003
为目标表面材料的红外发射率；
(42)在目标物方的辐射特性表达式的基础上考虑大气影响构建在目标像方的辐射特性表达式：
Figure PCTCN2017077101-appb-000004
ρθ为透过率，Lθ为大气程辐射。
进一步，所述目标为飞机，飞机的辐射亮度主要分为飞机蒙皮、发动机、尾焰三部分分别计算。
进一步，计算飞机蒙皮辐射亮度时首先对在位置(x,y,z)处的目标物方温度T(x,y,z)进行处理，得到考虑速度后的目标物方温度：
Figure PCTCN2017077101-appb-000005
式中，k为恢复系数，γ为空气的定压热容量和定容热容量之比，M为速度。
进一步，所述步骤(8)的具体实现方式为：首先根据步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线初步判定飞机姿态，从所述映射数据库中提取该姿态对应的光谱曲线集合；然后在光谱曲线集合中进行精确匹配，确定目标姿态。
进一步，所述目标为飞机，对步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线进行分析：
若以长波红外辐射为主，则初步判定姿态为迎头；
若短/中/长波辐射均存在，则初步判定姿态为侧面；
若短/中/长波段辐射均存在，且长波红外辐射最强，则初步判定姿态为俯视或仰视；
若短/中波最强，长波辐射最弱，则初步判定姿态为尾追。
一种点斑状目标姿态估计系统，包括离线训练部分和在线估计部分；
所述离线训练部分包括以下模块：
三维模型建立模块，用于建立目标三维几何模型并依据目标结构划分区域；
物方温度分布模型建立模块，用于针对目标的各区域建立物方温度分布模型；
红外辐射传输模型建立模块，用于建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模型；所述六种姿态包括顶视、俯视、迎头、追尾、左右、侧面；
像方辐射能量模型建立模块，用于利用已建的物方温度分布模型和红外辐射传输模型构建目标在六种典型姿态下的像方辐射能量模型；
映射数据库建立模块，用于分别在六种典型姿态下，结合目标像方辐射能量模型，仿真计算得到关于波长-目标像方辐射能量的点斑状目标红外光谱曲线，从而建立目标姿态与光谱的映射数据库；
所述在线估计部分包括以下模块：
探测模块，用于探测点斑状目标；
点斑状目标红外光谱曲线生成模块，用于实时采集点斑状目标像方辐射能量，绘制关于波长-目标像方辐射能量的点斑状目标红外光谱曲线；
姿态估计模块，用于将生成的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线在目标姿态与光谱的映射数据库内进行匹配，确定目标姿态。
进一步，所述像方辐射能量模型建立模块包括：
目标物方的辐射特性表达式构建子模块，根据物方温度分布模型构建目标物方的辐射特性表达式L(T(x,y,z))：
已知在波段λ1～λ2、温度为T的等效黑体辐射亮度Lb(T(x,y,x))为：
Figure PCTCN2017077101-appb-000006
则目标表面红外辐射亮度L(T(x,y,z))为：
Figure PCTCN2017077101-appb-000007
式中：λ为波长，T(x,y,z)为在位置(x,y,z)处的目标物方温度，黑体发射率ε＝1，
Figure PCTCN2017077101-appb-000008
为目标表面材料的红外发射率；
目标像方的辐射特性表达式构建子模块，用于在目标物方的辐射特性表达式的基础上考虑大气影响构建在目标像方的辐射特性表达式：
Figure PCTCN2017077101-appb-000009
ρθ为透过率，Lθ为大气程辐射。
进一步，所述姿态估计模块包括：
初步判定子模块，用于根据生成的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线初步判定飞机姿态，从所述映射数据库中提取该姿态对应的光谱曲线集合；
精确匹配子模块，用于在光谱曲线集合中进行精确匹配，确定目标姿态。
本发明的有益技术效果体现在：
(1)将目标姿态有效划分为6种典型姿态，简单合理且具代表性。目标在空间分布上是三维的，在不同的观测角下可体现出目标的多种姿态信息。为了简化处理，将目标的姿态划分为迎头、尾追、仰视、俯视、左右侧视这6种姿态，然后建立与这6种姿态对应的光谱数据库。
(2)提出目标的姿态与观察所得的光谱曲线形状有关。目标的红外辐射光谱主要是由温度决定，而目标自身的材质结构及组成对目标表面各点处的温度有着重要影响。即目标的温度可根据结构划分为几个不同区域，而目标在不同姿态下所观察到的目标区域也不相同，从而使得观测到的点斑状目标的光谱曲线也不相同。
(3)利用三维目标的红外光谱形态估计目标当前姿态。分析比较测到的点斑状目标的红外光谱各波段内辐射能量值的差异信息，结合因目标结构而形成的温度分布及6种典型姿态下的光谱特性估计该测量的点目标当前所处的姿态。
总的来说，本发明考虑了目标随空间位置、姿态、探测距离等多种因素变化的温度分布及辐射特性；将目标的姿态进行了有效的区分，该区分方式简单合理且准确性高，解决了如何根据红外光谱信息估计点斑状目标姿态的难题；
[附图说明]
图1为点斑状目标姿态估计方法流程图；
图2为飞机光谱特征组成示意图；
图3为飞机与天空、地球、太阳等平均光谱特性曲线图；
图4为不同方位角下的飞机光谱数据示意图；
图5为某姿态下的SU-27辐射光谱图；
图6为大气层内探测系统观测大气层内目标时相对位置关系示意图。
[具体实施方式]
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。此外，下面所描述的本发明各个实施方式中所涉及到的技术特征只要彼此之间未构成冲突就可以相互组合。
一种点斑状目标姿态估计方法包括离线训练部分和在线评估部分。离线训练部分包括建立目标的三维几何模型、温度分布模型及大气传输模型；设定仿真参数，仿真计算6种姿态下点斑状目标的红外辐射光谱，建立“姿态-光谱”数据库；在线评估部分包括分析测得的红外光谱形态，比对“姿态-光谱”数据库估计该目标的当前姿态。
本发明方法流程如图1所示，以某型飞机为例，具体的实施方法包括以下步骤：
一、离线训练步骤
(1)建立目标三维几何模型并依据目标结构划分区域。
本实例中，根据搜查到的某型飞机自身的尺寸及结构信息，利用3dsmax软件建立飞机的三维模型，并正确区分飞机的不同部位。飞机部位主要包括尾焰、发动机、机翼前缘、机翼后缘、机头、座舱级一般部位这几个部分。
(2)针对目标的各区域建立物方温度分布模型。
针对目标的各区域建立物方温度分布模型，可实时或仿真测量，也可采用拟合方法。本发明优选采用插值拟合法，具体而言，根据已查到的飞机温度数据，针对不同区域对其温度数据进行插值拟合处理，从而近似得到各区域内随空间位置变化的温度分布函数，然后根据拟合得到的函数计算各区域内其它位置的目标点的温度。
(3)建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模型。
通过调整目标与探测系统间的位置可形成6种姿态，包括顶视、俯视、迎头、追尾、左右、侧面。分别建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模，建立过程如下：
大气层内探测系统观测大气层内目标条件下的辐射传输路径如图5所示：大气层内的探测系统的位置由A点表示，被观测的大气层内目标的位置由B点表示，被观测目标的辐射能量由B点向A点传播。
结合大气传输软件计算大气层内探测系统到空间目标的大气红外透过率和程辐射时其输入参数为：探测系统高度、目标高度、探测系统与目标间的斜距、大气层最高高度、天顶角、地球半径、红外波段。其中探测系统高度和红外波段依据实际情况确定；地球半径为探测系统所处纬度处的半径值；大气层最高高度设定为100公里；天顶角如图6中的∠CAB所示，其计算公式为：
Figure PCTCN2017077101-appb-000010
其中
Figure PCTCN2017077101-appb-000011
分别为探测系统和被观测目标距地心的高度，
Figure PCTCN2017077101-appb-000012
为探 测系统和被观测目标之间的距离。
输入以上参数即可获得大气层内探测系统观测目标在角度θ下的透过率ρθ和程辐射Lθ。
例如，经计算：在正下方(θ＝0)高度50km的探测系统对高度10km且距离其40km的空间目标探测时计算得到的在8～12μm的透过率ρθ＝0＝0.9977和程辐射Lθ＝0＝0.001905(W·m-2·sr-1)。
(4)利用已建的物方温度分布模型和红外辐射传输模型构建目标在六种典型姿态下的像方辐射能量模型。
(41)根据物方温度分布模型构建目标物方的辐射特性表达式L(T(x,y,z))。
已知在波段λ1～λ2、温度为T(x,y,z)的等效黑体(发射率ε＝1)辐射亮度Lb(T(x,y,x))为：
Figure PCTCN2017077101-appb-000013
其中：Lb(T(x,y,z))——等效黑体红外辐射亮度(W·m-2·sr-1)；
λ——波长(μm)；
T(x,y,z)——在位置(x,y,z)处的目标物方温度(K)。
因此，表面温度为T的目标表面红外辐射亮度L(T)的计算公式为：
Figure PCTCN2017077101-appb-000014
式中：
L(T(x,y,z))：目标表面红外辐射亮度(W·m-2·sr-1)；
Lb(T(x,y,z))：温度为T(x,y,z)的黑体红外辐射亮度(W·m-2·sr-1)；
Figure PCTCN2017077101-appb-000015
目标表面材料的红外发射率。
(42)在目标物方的辐射特性表达式的基础上考虑大气影响构建在目标像方的辐射特性表达式L(T(x,y,z),θ)。
Figure PCTCN2017077101-appb-000016
在本发明实例中，飞机的红外辐射组成如图2和图3所示，因此主要计算飞机蒙皮、发动机、尾焰等部位的辐亮度。具体计算如下：
(a)发动机辐射计算
取飞机发动机尾喷管的发射率
Figure PCTCN2017077101-appb-000017
然后根据公式L(T(x,y,z),θ)计算发动机尾喷管处的辐射亮度。
(b)飞机蒙皮辐射计算
飞机表面温度同飞行速度(马赫数M)有关，首先要对T(x,y,z)进行考虑速度因素的处理，然后利用公式L(T(x,y,z),θ)计算蒙皮处的辐射亮度。
对于在对流层中飞行速度不是很高(M≤2.5)的飞机，飞机表面温度T2(x,y,z)表达为：
Figure PCTCN2017077101-appb-000018
式中:T2(x,y,z)——考虑速度的飞机蒙皮表面温度(K)；
k——恢复系数，通常在0.8～0.94，对于层流取k＝0.82.；
γ——空气的定压热容量和定容热容量之比，γ＝1.4；
M——速度即飞行马赫数。
然后按公式L(T2(x,y,z),θ)计算飞机蒙皮的辐射亮度值，蒙皮发射率
Figure PCTCN2017077101-appb-000019
为0.85。
(c)飞机尾焰辐射计算
取飞机发动机尾喷管的发射率
Figure PCTCN2017077101-appb-000020
然后根据公式L(T(x,y,z),θ)计算发动机尾喷管处的辐射亮度。
(5)分别在六种典型姿态下，结合目标像方辐射能量模型，仿真计算得到关于波长-目标像方辐射能量的点斑状目标红外光谱曲线，从而建立目标姿态与光谱的映射数据库。本发明实例的部分数据如图4所示。
二、在线估计步骤：
(6)探测点斑状目标
设定探测系统的光学属性、光谱分辨率、成谱波段，动目标的运动速度、飞行高度以及探测系统与动目标间的距离等仿真参数，启动探测,得到点斑状目标图像。
(7)实时采集点斑状目标像方辐射能量，绘制关于波长-目标像方辐射能量的点斑状目标红外光谱曲线。
从点斑状目标图像获取像素灰度信息，通过定标方式获取其对应的像方辐射能量，根据多点的像方辐射能量点斑状目标红外光谱曲线。
(8)将所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线在目标姿态与光谱的映射数据库内进行匹配，确定目标姿态。
首先，按照下面典型姿态光谱特性初步判定飞机姿态，从数据库中提取该姿态对应的光谱曲线集合。
①迎头方向姿态：从飞机机头往尾部方向观测，此时形成的点目标的光谱以蒙皮的长波红外辐射为主，其它分量较少。
②侧面方向姿态：当飞机对称时，左右侧面的光谱基本一致，此时生成的点目标的光谱包括飞机蒙皮、尾喷管以及尾焰的辐射，且短/中/长波辐射均有，各分量齐全。
③俯视或仰视方向姿态：从飞机的正上方向下或正下方向向上观看，观测到的蒙皮面积最大，生成的光谱中短/中/长波段的辐射均存 在，且长波红外辐射最强。
④尾追方向姿态：从飞机的尾部往机头方向观测，此时机头观测不到且观测到的蒙皮面积小，主要辐射为尾焰和喷管，短/中波最强，长波辐射最少。
然后，在初步判定得到的光谱曲线集合，对点斑状目标红外光谱曲线进行精确匹配，确定目标姿态。
如图5所示为实际测量到的SU-27型飞机的红外辐射光谱，经过定性分析和定量计算该光谱可知在短波、中波、长波均有辐射分量，而中长波5.5μm～7.5μm内无辐射，且中波辐射最强，短波次之，长波辐射最弱。通过匹配数据库中的典型姿态的光谱特性可知此时测量到的SU-27飞机处于尾追方向姿态，观测到的主要是尾焰和尾喷管，而尾焰和尾喷管的温度是飞机表面所有点的温度中较高的，也符合理论分析。

Claims (9)

  1. 一种点斑状目标姿态估计方法，其特征在于，包括离线训练部分和在线估计部分；
    所述离线训练部分包括以下步骤：
    (1)建立目标三维几何模型并依据目标结构划分区域；
    (2)针对目标的各区域建立物方温度分布模型；
    (3)建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模型；所述六种姿态包括顶视、俯视、迎头、追尾、左右、侧面；
    (4)利用已建的物方温度分布模型和红外辐射传输模型构建目标在六种典型姿态下的像方辐射能量模型；
    (5)分别在六种典型姿态下，结合目标像方辐射能量模型，仿真计算得到关于波长-目标像方辐射能量的点斑状目标红外光谱曲线，从而建立目标姿态与光谱的映射数据库；
    所述在线估计部分包括以下步骤：
    (6)探测点斑状目标；
    (7)实时采集点斑状目标像方辐射能量，绘制关于波长-目标像方辐射能量的点斑状目标红外光谱曲线；
    (8)将步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线在目标姿态与光谱的映射数据库内进行匹配，确定目标姿态。
  2. 根据权利要求1所述的点斑状目标姿态估计方法，其特征在于，所述步骤(4)的具体实现过程为：
    (41)根据物方温度分布模型构建目标物方的辐射特性表达式L(T(x,y,z))：
    已知在波段λ1～λ2、温度为T的等效黑体辐射亮度Lb(T(x,y,x))为：
    Figure PCTCN2017077101-appb-100001
    则目标表面红外辐射亮度L(T(x,y,z))为：
    Figure PCTCN2017077101-appb-100002
    式中：λ为波长，T(x,y,z)为在位置(x,y,z)处的目标物方温度，黑体发射率ε＝1，
    Figure PCTCN2017077101-appb-100003
    为目标表面材料的红外发射率；
    (42)在目标物方的辐射特性表达式的基础上考虑大气影响构建在目标像方的辐射特性表达式：
    Figure PCTCN2017077101-appb-100004
    ρθ为透过率，Lθ为大气程辐射。
  3. 根据权利要求2所述的点斑状目标姿态估计方法，其特征在于，所述目标为飞机，飞机的辐射亮度分为飞机蒙皮、发动机、尾焰三部分分别计算。
  4. 根据权利要求3所述的点斑状目标姿态估计方法，其特征在于，计算飞机蒙皮辐射亮度时首先对在位置(x,y,z)处的目标物方温度T(x,y,z)进行处理，得到考虑速度后的目标物方温度：
    Figure PCTCN2017077101-appb-100005
    式中，k为恢复系数，γ为空气的定压热容量和定容热容量之比，M为速度。
  5. 根据权利要求1所述的点斑状目标姿态估计方法，其特征在于， 所述步骤(8)的具体实现方式为：首先根据步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线初步判定飞机姿态，从所述映射数据库中提取该姿态对应的光谱曲线集合；然后在光谱曲线集合中进行精确匹配，确定目标姿态。
  6. 根据权利要求1所述的点斑状目标姿态估计方法，其特征在于，所述目标为飞机，对步骤(7)所得的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线进行分析：
    若以长波红外辐射为主，则初步判定姿态为迎头；
    若短/中/长波辐射均存在，则初步判定姿态为侧面；
    若短/中/长波段辐射均存在，且长波红外辐射最强，则初步判定姿态为俯视或仰视；
    若短/中波最强，长波辐射最弱，则初步判定姿态为尾追。
  7. 一种点斑状目标姿态估计系统，其特征在于，包括离线训练部分和在线估计部分；
    所述离线训练部分包括以下模块：
    三维模型建立模块，用于建立目标三维几何模型并依据目标结构划分区域；
    物方温度分布模型建立模块，用于针对目标的各区域建立物方温度分布模型；
    红外辐射传输模型建立模块，用于建立探测系统观测大气层内目标在六种典型姿态下的红外辐射传输模型；所述六种姿态包括顶视、俯视、迎头、追尾、左右、侧面；
    像方辐射能量模型建立模块，用于利用已建的物方温度分布模型和红外辐射传输模型构建目标在六种典型姿态下的像方辐射能量模型；
    映射数据库建立模块，用于分别在六种典型姿态下，结合目标像方辐射能量模型，仿真计算得到关于波长-目标像方辐射能量的点斑状目标红外光谱曲线，从而建立目标姿态与光谱的映射数据库；
    所述在线估计部分包括以下模块：
    探测模块，用于探测点斑状目标；
    点斑状目标红外光谱曲线生成模块，用于实时采集点斑状目标像方辐射能量，绘制关于波长-目标像方辐射能量的点斑状目标红外光谱曲线；
    姿态估计模块，用于将生成的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线在目标姿态与光谱的映射数据库内进行匹配，确定目标姿态。
  8. 根据权利要求7所述的点斑状目标姿态估计系统，其特征在于，所述像方辐射能量模型建立模块包括：
    目标物方的辐射特性表达式构建子模块，根据物方温度分布模型构建目标物方的辐射特性表达式L(T(x,y,z))：
    已知在波段λ1～λ2、温度为T的等效黑体辐射亮度Lb(T(x,y,x))为：
    Figure PCTCN2017077101-appb-100006
    则目标表面红外辐射亮度L(T(x,y,z))为：
    Figure PCTCN2017077101-appb-100007
    式中：λ为波长，T(x,y,z)为在位置(x,y,z)处的目标物方温度，黑体发射率ε＝1，
    Figure PCTCN2017077101-appb-100008
    为目标表面材料的红外发射率；
    目标像方的辐射特性表达式构建子模块，用于在目标物方的辐射特性表达式的基础上考虑大气影响构建在目标像方的辐射特性表达式：
    Figure PCTCN2017077101-appb-100009
    ρθ为透过率，Lθ为大气程辐射。
  9. 根据权利要求1所述的点斑状目标姿态估计方法，其特征在于，所述姿态估计模块包括：
    初步判定子模块，用于根据生成的关于波长-目标像方辐射能量的点斑状目标红外光谱曲线初步判定飞机姿态，从所述映射数据库中提取该姿态对应的光谱曲线集合；
    精确匹配子模块，用于在光谱曲线集合中进行精确匹配，确定目标姿态。
PCT/CN2017/077101 2016-12-31 2017-03-17 一种点斑状目标姿态估计方法及系统 WO2018120443A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/458,221 US11162847B2 (en) 2016-12-31 2019-07-01 Method and system of attitude estimation of spotted target

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201611268809.8A CN106706133B (zh) 2016-12-31 2016-12-31 一种点斑状目标姿态估计方法及系统
CN201611268809.8 2016-12-31

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/458,221 Continuation-In-Part US11162847B2 (en) 2016-12-31 2019-07-01 Method and system of attitude estimation of spotted target

Publications (1)

Publication Number Publication Date
WO2018120443A1 true WO2018120443A1 (zh) 2018-07-05

Family

ID=58906558

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/077101 WO2018120443A1 (zh) 2016-12-31 2017-03-17 一种点斑状目标姿态估计方法及系统

Country Status (3)

Country Link
US (1) US11162847B2 (zh)
CN (1) CN106706133B (zh)
WO (1) WO2018120443A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117576362A (zh) * 2024-01-16 2024-02-20 国科大杭州高等研究院 一种基于遮挡比的低分辨率热红外影像飞机识别方法

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109446559A (zh) * 2018-09-20 2019-03-08 中国人民解放军63653部队 一种浮空器辐射热环境特性估算方法
CN110807270B (zh) * 2019-11-13 2023-09-29 北京环境特性研究所 一种基于尾焰辐射线型反演发动机参数及预估弹道的方法
CN110888121B (zh) * 2019-12-16 2020-11-24 上海智瞳通科技有限公司 目标体探测方法与装置及目标体温度探测方法与装置
CN111829666B (zh) * 2020-07-17 2022-12-06 中国人民解放军火箭军工程大学 一种目标红外成像仿真模型的四级验证方法
CN113092368B (zh) * 2021-03-16 2022-12-13 上海机电工程研究所 一种基于无人机的红外波段大气透过率测量方法及系统
CN116740507A (zh) * 2023-08-02 2023-09-12 中科星图测控技术股份有限公司 一种基于arm架构的空间目标检测模型构建方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030155501A1 (en) * 2002-02-15 2003-08-21 International Business Machines Corporation Method and apparatus for compensating waveforms, spectra, and profiles derived therefrom for effects of drift
CN104121992A (zh) * 2014-07-10 2014-10-29 上海宇航系统工程研究所 空间目标红外辐射全任务空域动态特性分析方法
CN105243289A (zh) * 2015-11-17 2016-01-13 上海无线电设备研究所 复杂目标红外辐射特性一体化建模方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013525195A (ja) * 2010-04-29 2013-06-20 ノルスク・インスティテュット・フォル・ルフトフォルスクニング 航空機の前方の悪大気状態を検出するためのシステムおよび方法
CN101976275B (zh) * 2010-09-21 2012-11-07 北京航空航天大学 飞机红外辐射与大气透过率建模方法
CN102270355A (zh) * 2011-04-28 2011-12-07 华中科技大学 一种基于景物分类的红外场景图像生成方法
CN104501959B (zh) * 2014-12-30 2016-08-17 华中科技大学 一种红外图谱关联智能探测方法及装置
US9963251B2 (en) * 2015-05-27 2018-05-08 The Aerospace Corporation Systems and methods for estimating parameters of a spacecraft based on emission from an atomic or molecular product of a plume from the spacecraft
CN105678230B (zh) * 2015-12-30 2017-04-05 华中科技大学 一种红外目标投影模型指导的飞机感兴趣区测谱方法
CN105956591B (zh) * 2016-04-28 2019-03-29 中山大学 一种在线高温零件红外图像光谱抽样检测方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030155501A1 (en) * 2002-02-15 2003-08-21 International Business Machines Corporation Method and apparatus for compensating waveforms, spectra, and profiles derived therefrom for effects of drift
CN104121992A (zh) * 2014-07-10 2014-10-29 上海宇航系统工程研究所 空间目标红外辐射全任务空域动态特性分析方法
CN105243289A (zh) * 2015-11-17 2016-01-13 上海无线电设备研究所 复杂目标红外辐射特性一体化建模方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHAI SHIJIE ET AL.: "Infrared Radiation Characteristics and Detection of Stealth Aeroplane, Fire Control & Command Control", 25 August 2014 (2014-08-25), pages 26 - 29, ISSN: 1002-0640 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117576362A (zh) * 2024-01-16 2024-02-20 国科大杭州高等研究院 一种基于遮挡比的低分辨率热红外影像飞机识别方法
CN117576362B (zh) * 2024-01-16 2024-05-24 国科大杭州高等研究院 一种基于遮挡比的低分辨率热红外影像飞机识别方法

Also Published As

Publication number Publication date
CN106706133B (zh) 2017-09-29
US11162847B2 (en) 2021-11-02
US20190323894A1 (en) 2019-10-24
CN106706133A (zh) 2017-05-24

Similar Documents

Publication Publication Date Title
WO2018120443A1 (zh) 一种点斑状目标姿态估计方法及系统
WO2018120444A1 (zh) 一种动目标的红外辐射光谱特性仿真分析方法
CN107514993B (zh) 基于无人机的面向单体建筑建模的数据采集方法及系统
CN111028220B (zh) 一种点云铆钉齐平度自动检测方法
WO2016106961A1 (zh) 一种多传感器融合的超近距离自主导航装置与方法
CN106651953A (zh) 一种基于交通指示牌的车辆位姿估计方法
CN103557792B (zh) 一种空中加油锥套目标的视觉跟踪与位置测量方法
CN106772417B (zh) 一种动目标多维度多尺度红外光谱特征测量方法及系统
CN105716539A (zh) 一种快速高精度的三维形面测量方法
CN106950981A (zh) 一种无人机高度控制方法及系统
CN111241667B (zh) 一种基于图像处理和探针数据处理识别等离子体位形方法
CN109118534B (zh) 实时获取冰风洞内模型表面上结冰冰型详细信息的方法
CN107368617B (zh) 基于Lowtran7大气软件的地空探测红外成像系统作用距离计算方法
CN115731268A (zh) 基于视觉/毫米波雷达信息融合的无人机多目标跟踪方法
CN107607091A (zh) 一种测量无人机飞行航迹的方法
CN103942786B (zh) 无人机可见光和红外图像的自适应团块目标检测方法
CN105160341B (zh) 一种基于直线段检测和内心性质的三角形检测方法
CN113838010A (zh) 基于红外热波及卷积神经网络的复合材料损伤智能检测方法
CN115272458A (zh) 用于固定翼无人机着陆阶段的视觉定位方法
CN113642267B (zh) 一种飞行器表面前缘附着线区域提取方法
CN104075710B (zh) 一种基于航迹预测的机动扩展目标轴向姿态实时估计方法
CN113740864A (zh) 基于激光三维点云的探测器软着陆末段自主位姿估计方法
CN112991433A (zh) 基于双目深度感知和车辆位置的货车外廓尺寸测量方法
CN105678230B (zh) 一种红外目标投影模型指导的飞机感兴趣区测谱方法
Zhang et al. A binocular vision based auxiliary navigation system of the unmanned aerial vehicle for power line inspection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17889230

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17889230

Country of ref document: EP

Kind code of ref document: A1