CN105962904A - Human tissue focus detection method based on infrared thermal imaging technology - Google Patents

Abstract

The invention discloses a human tissue focus detection method based on an infrared thermal imaging technology. Specifically, the detection method is implemented by the following steps: 1, establishing an infrared temperature measuring physical model; 2, in accordance with a corresponding relation between temperature analyzed by the physical model and obtained in the step and a gray scale, converting an image obtained by virtue of an infrared thermal imaging instrument into a gray scale image; 3, processing the gray scale image obtained in the step 2; and 4, contrasting and analyzing a processing result obtained in the step 3 and a medical image, so as to discover a human tissue focus as quickly as possible. With the application of the detection method disclosed by the invention, the problem in the prior art that a human tissue focus, which is late in detection time and large in detection volume, can be detected out when the focus has developed into a certain density and a certain volume can be avoided.

Description

A kind of tissue focus detection method based on Infrared Thermography Technology

Technical field

The invention belongs to technical field of medical image processing, be specifically related to a kind of tissue based on Infrared Thermography Technology Focus detection method.

Background technology

Medical infrared thermal imaging system is for detecting human infrared radiation, union optics, precision instrument, thermal imaging, electricity Son is learned, image processing techniques high new instrument, by thermographic form display human skin's Temperature Distribution Change, variation portion and degree.The thermography that medical worker obtains by observing thermal infrared imager quantitative analysis can measure temperature And combine clinical experience, focus is diagnosed.

Owing to Infrared Thermography Technology can obtain real time information when cell tissue metabolism changes, thus, every The disease that can cause tissue thermal change all can be examined by this technology, and its clinical value is high, applied range.Can lead to Cross the infrared emanation signal during thermal infrared imager passively receives human metabolism, measure the number of infrared source Value, form and the degree of depth, respectively organize and the corresponding relation of disease Yu the thermal-radiating distribution of whole body, through arrangement amount according to human body The infrared thermal imagery formed after change is the distributed image of human epidermal temperature.Human body as a natural biological heater, by In tissue metabolism, anatomical structure, neural state and sanguimotor difference, the temperature at each position of body also differs, and continues And define different Temperature Distribution thermal fields.According to the heat radiation difference of exception Yu normal tissue regions, the most permissible in conjunction with clinic Diagnose and infer the nature and extent of disease.

Present stage medical pathology diagnosis mainly uses the structures such as nuclear magnetic resonance, NMR (MRI), X-ray, tomoscan (CT), B ultrasonic The technical method of image, but need to show that when focus is developed to certain density, certain volume this abnormal structure becomes Change.Infrared thermal imaging technique is applied and is then belonged to functional imaging on medical imaging, from disease occur, development typically enter From the point of view of journey, the change of organ dysfunction or tissue characteristics is often prior to its form and the variation of structure.When shell temperature change reaches During the resolution of thermal imaging system, medical infrared thermal imaging system just can detect and recorded this change, and show abnormality high temperature or low The position of temperature, it is possible to realize the early discovery of tissue focus, in order to early diagnosis, treat as early as possible, simultaneously accurate quick, real With efficiently, and tissue be there is no any damage.

Summary of the invention

It is an object of the invention to provide a kind of tissue focus detection method based on Infrared Thermography Technology, solve existing There are present in technology tissue focus detection evening time, detection volume big and cause focus to be developed to certain density, certain The problem being just detected during volume.

The technical solution adopted in the present invention is, a kind of tissue focus detection method based on Infrared Thermography Technology, Specifically implement according to following steps:

Step 1, set up infrared measurement of temperature physical model；

Step 2, utilize physical model analysis temperature and the corresponding relation of gray scale that step 1 obtains, infrared thermal imagery will be passed through The temperature pattern that instrument obtains is converted into gray level image；

Step 3, the gray level image obtaining step 2 process；

Step 4, result step 3 obtained are analyzed with medical image, and finder soma is sick early Stove.

The feature of the present invention also resides in,

Step 1 is specifically implemented according to following steps:

Step (1.1), first trying to achieve radiant illumination according to infrared radiation temperature formula, formula is as follows:

E_λ=A₀d^-2[τ_aλε_λL_bλ(T₀)+τ_aλ(1-α_λ)L_bλ(T_u)+ε_aλL_bλ(T_a)] (1)

(1) in formula, α_λFor testee Surface absorption rate, ε_λFor testee slin emissivity, ε_aλFor atmospheric emission rate, τ_aλFor the spectral-transmission favtor of air, T₀For object surface temperature, T_aFor atmospheric temperature, T_uFor ambient temperature, d is measured object Surface is to the vertical dimension between measuring instrument, A₀d^-2For constant, A₀Effective area for target；

Step (1.2), the radiant illumination obtained according to step (1.1) try to achieve the response voltage of thermal infrared imager:

Generally infrared thermography is operated in a wavelength band the narrowest, between 8～14 μm or between 3～5 μm, Then think the ε in (1) formula_λ、α_λ、τ_aλUnrelated with λ, then obtain the response voltage V of thermal infrared imager_SFor:

$V_S=A_R{A}_0{d}^{-2}\{{\mathrm{\τ}}_a\[\mathrm{\ϵ}\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_0\right)d\mathrm{\λ}+(1-a)\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_u\right)d\mathrm{\λ}\]+{\mathrm{\ϵ}}_a\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_a\right)d\mathrm{\λ}\}---\left(2\right)$

(2) in formula, A_RFor the area of thermal imaging system lens, A₀For the effective area of target, A₀d^-2For constant, make K=A_RA₀d^-2,Then (2) formula becomes

V_S=K{ τ_a[εf(T₀)+(1-a)f(T_u)+ε_af(T_a)]} (3)

Step (1.3), calculating body surface true temperature:

According to Planck blackbody radiation law, can obtain

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-a)T_u^n\]+{\mathrm{\ϵ}}_a{T}_a^n---\left(4\right)$

(4) in formula, T_rFor the radiation temperature on testee surface, T_oFor the true temperature on testee surface, T_uFor environment Temperature, T_aFor atmospheric temperature, α is testee Surface absorption rate, and ε is surface emissivity, ε_aFor atmospheric emission rate, τ_aFor The absorbance of air, the value of n is different with wave band difference when using thermal imaging system, and to InSb (3～5 μm) detector, n value is 8.68, to HgCdTe (6～9 μm) detector, n value is 5.33, and to HgCdTe (8～14 μm) detector, n value is 4.09；

If testee surface true temperature is T₀, then T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-a)T_u^n-\frac{{\mathrm{\ϵ}}_a}{{\mathrm{\τ}}_a}{T}_a^n\]\}}^{\frac{1}{n}}---\left(5\right)$

Step (1.4), when human body surface emissivity keep constant, human body can be seen as grey body, when testee surface When meeting grey body approximation, i.e. ε=α, and if thinking air ε_α=α_a=1-τ_a, then the true temperature of ash surface is calculated as follows:

By above-mentioned (3) Shi Ke get:

V_S=K{ τ_a[εf(T₀)+(1-ε)f(T_u)]+(1-τ_a)f(T_a)} (6)

By above-mentioned (4) Shi Ke get:

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-\mathrm{\ϵ})T_u^n+(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]---\left(7\right)$

Can be obtained grey body surface true temperature by above-mentioned (5) formula is T₀, T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-\mathrm{\ϵ})T_u^n-(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]\}}^{\frac{1}{n}}---\left(8\right)$

Step (1.5), calculate 2 true temperatures of testee difference particularly as follows:

Closely during thermometric, ignore the impact of atmospheric transmittance, make τ_a=1, the formula (7) in the most described step 1.4 becomes

$T_r=T_0{\{\mathrm{\ϵ}\[1-{\left(\frac{T_u}{T_0}\right)}^n\]+{\left(\frac{T_u}{T_0}\right)}^n\}}^{\frac{1}{n}}---\left(9\right)$

Formula (8) becomes

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[T_r^n-(1-\mathrm{\ϵ})T_u^n\]\}}^{\frac{1}{n}}---\left(10\right)$

Obtain the true temperature of body surface,

When object surface temperature is much larger than ambient temperature, i.e. T_u/T₀＞ ＞ 0, then (9), (10) formula become

$T_r={\mathrm{\ϵ}}^{\frac{1}{n}}{T}_0---\left(11\right)$

$T_0=\frac{T_r}{\sqrt[n]{\mathrm{\ϵ}}}---\left(12\right)$

If the radiation temperature measuring testee surface any two points is T respectively_r1And T_r2, then this true temperature of 2 Difference is

$\mathrm{\Δ}T=T_{01}-T_{02}=\frac{1}{{\mathrm{\ϵ}}^{1/n}}\{{\[T_{r1}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}-{\[T_{r2}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}\}---\left(13\right)$

Step 2 particularly as follows:

Step (2.1), when in step 1 between any two points of testee surface shell temperature change reach thermal infrared imager Resolution time, use thermal infrared imager detection and recorded this change, the position of show abnormality high temperature or low temperature, aobvious Demonstrate the temperature distribution image of body surface with the form of pseudo-colours in display screen.

Step (2.2), use high-precision blackbody furnace as standard before test object surface temperature, draw out photoelectricity Switching device output signal and the relation curve of blackbody furnace temperature, the relation right and wrong between emittance and the temperature of black body emission Linear, it is calculated by thermal imaging system spectral response and Planck's law of radiation, in order to set up between amount of radiation and temperature Relation, carries out different temperatures and arranges and measure it, and carried out by accurate to measurement result and black matrix temperature value black matrix Matching, has just obtained calibration curve, by the sensor on camera lens, the infrared energy of testee is converted into the signal of telecommunication, These raw electrical signals are processed and are transformed into gray value by subsequent conditioning circuit after camera lens further again, demonstrate at display Come, through such a process, just obtained the corresponding relation between image intensity value and object temperature, and then obtained gray-scale map Picture.

Step 3, specifically, Medical Image Processing algorithm research, is being gone based on the wavelet package transforms image revising Wiener filtering Make an uproar, small echo self-adaptive spot noise based on Pulse-coupled Neural Network Model filters, based on medical image increasings such as wavelet transformations The rim detection new algorithm that noise is more sane it is proposed on the basis of strong algorithms.It is based on Corner Detection sane to noise to design Edge detector, utilize multiple dimensioned anisotropic Gaussian direction gradient wave filter to extract edge and characteristic point structural information:

Step (3.1), set up different types of edge model, including staged, pulsed, ramp type and double ridge edge Model, and the model of characteristic point, including L-type, Y type, X-type, the model of characteristic point of star-like, end type, notch cuttype, then set up Can effectively distinguish the anisotropic Gaussian direction gradient relative to different local feature models of adjacent edge and characteristic point Expression formula, and the measure of feature degree of isolation, carry out the measurement of noise sensitivity and estimation and in geometry and gray scale The tolerance of robustness in conversion；

Step (3.2), link anisotropic Gaussian direction gradient represent；

Step (3.3), infrared image is carried out logarithmic transformation, make speckle noise be converted to additive noise, by infrared image Process through Wiener filtering, calculate the standard variance of its additive noise, in this, as wavelet threshold, then, utilize wavelet transformation pair Image carries out pretreatment, utilizes PCNN to revise wavelet coefficient accordingly in wavelet field, finally, carries out wavelet inverse transformation And exponential transform, it is thus achieved that filter the image of noise.

The invention has the beneficial effects as follows, a kind of tissue focus detection method based on Infrared Thermography Technology, according to red Outer temperature-measurement principle and heat radiation theorem, the triangular inherent pass of further investigation human infrared radiation-surface temperature-substance characteristics System, sets up medical science infrared imaging physical model, medical science infrared diagnostics Image semantic classification system, medical science based on Infrared Thermography Technology Image procossing and diagnostic platform, then combine with structure image technology, many pathological changes can be made to obtain early discovery, disease rule obtains To more comprehensively recognizing, disease character obtains more Accurate Diagnosis.Compared to the structure image technology such as nuclear magnetic resonance, NMR, B ultrasonic, energy of the present invention Enough realize the discovery as early as possible of tissue focus, accurate quick, practicality and high efficiency, and tissue is not had any damage.

Accompanying drawing explanation

Fig. 1 (a) is to adopt in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention Experiment original image；

Fig. 1 (b) is to add in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention The speckle noise image added；

Fig. 1 (c) is warp in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention Image after allusion quotation denoising method Wiener filtering (Wiener) process；

Fig. 1 (d) is base in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention Image after the median filter method (PCNN-MF) of PCNN processes；

Fig. 1 (e) is base in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention Image after the wavelet soft-threshold filter method (PCNN-WD) of PCNN processes；

Fig. 1 (f) is to adopt in a kind of tissue focus detection method experiment simulation based on Infrared Thermography Technology of the present invention Image after processing by the method for present invention proposition；

Fig. 2 is to use in a kind of tissue focus detection method validation verification based on Infrared Thermography Technology of the present invention Original image；

Fig. 3 is noisy in a kind of tissue focus detection method validation verification based on Infrared Thermography Technology of the present invention Image；

Fig. 4 is wiener in a kind of tissue focus detection method validation verification based on Infrared Thermography Technology of the present invention Image after filtering and noise reduction；

Fig. 5 is small echo in a kind of tissue focus detection method validation verification based on Infrared Thermography Technology of the present invention Image after bag denoising；

Fig. 6 is to revise in a kind of tissue focus detection method validation verification based on Infrared Thermography Technology of the present invention Image after Wiener filtering+wavelet packet denoising.

Detailed description of the invention

The present invention is described in detail with detailed description of the invention below in conjunction with the accompanying drawings.

A kind of tissue focus detection method based on Infrared Thermography Technology of the present invention, specifically real according to following steps Execute:

Step 1, set up infrared measurement of temperature physical model；

Step 2, utilize physical model analysis temperature and the corresponding relation of gray scale that step 1 obtains, infrared thermal imagery will be passed through The temperature pattern that instrument obtains is converted into gray level image；

Step 3, the gray level image obtaining step 2 process；

Step 4, result step 3 obtained are analyzed with medical image, and finder soma is sick early Stove.

Wherein, step 1 is specifically implemented according to following steps:

Step (1.1), first trying to achieve radiant illumination according to infrared radiation temperature formula, formula is as follows:

E_λ=A₀d^-2[τ_aλε_λL_bλ(T₀)+τ_aλ(1-α_λ)L_bλ(T_u)+ε_aλL_bλ(T_a)] (1)

(1) in formula, α_λFor testee Surface absorption rate, ε_λFor testee slin emissivity, ε_aλFor atmospheric emission rate, τ_aλFor the spectral-transmission favtor of air, T₀For object surface temperature, T_aFor atmospheric temperature, T_uFor ambient temperature, d is measured object Surface is to the vertical dimension between measuring instrument, A₀d^-2For constant, A₀Effective area for target；

Step (1.2), the radiant illumination obtained according to step (1.1) try to achieve the response voltage of thermal infrared imager:

Generally infrared thermography is operated in a wavelength band the narrowest, between 8～14 μm or between 3～5 μm, Then think the ε in (1) formula_λ、α_λ、τ_aλUnrelated with λ, then obtain the response voltage V of thermal infrared imager_SFor:

$V_S=A_R{A}_0{d}^{-2}\{{\mathrm{\τ}}_a\[\mathrm{\ϵ}\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_0\right)d\mathrm{\λ}+(1-a)\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_u\right)d\mathrm{\λ}\]+{\mathrm{\ϵ}}_a\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_a\right)d\mathrm{\λ}\}---\left(2\right)$

(2) in formula, A_RFor the area of thermal imaging system lens, A₀For the effective area of target, A₀d^-2For constant, make K=A_RA₀d^-2,Then (2) formula becomes

V_S=K{ τ_a[εf(T₀)+(1-a)f(T_u)+ε_af(T_a)]} (3)

Step (1.3), calculating body surface true temperature:

According to Planck blackbody radiation law, can obtain

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-a)T_u^n\]+{\mathrm{\ϵ}}_a{T}_a^n---\left(4\right)$

(4) in formula, T_rFor the radiation temperature on testee surface, T_oFor the true temperature on testee surface, T_uFor environment Temperature, T_aFor atmospheric temperature, α is testee Surface absorption rate, and ε is surface emissivity, ε_aFor atmospheric emission rate, τ_aFor The absorbance of air, the value of n is different with wave band difference when using thermal imaging system, and to InSb (3～5 μm) detector, n value is 8.68, to HgCdTe (6～9 μm) detector, n value is 5.33, and to HgCdTe (8～14 μm) detector, n value is 4.09；

If testee surface true temperature is T₀, then T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-a)T_u^n-\frac{{\mathrm{\ϵ}}_a}{{\mathrm{\τ}}_a}{T}_a^n\]\}}^{\frac{1}{n}}---\left(5\right)$

Step (1.4), when human body surface emissivity keep constant, human body can be seen as grey body, when testee surface When meeting grey body approximation, i.e. ε=α, and if thinking air ε_α=α_a=1-τ_a, then the true temperature of ash surface is calculated as follows:

By above-mentioned (3) Shi Ke get:

V_S=K{ τ_a[εf(T₀)+(1-ε)f(T_u)]+(1-τ_a)f(T_a)} (6)

By above-mentioned (4) Shi Ke get:

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-\mathrm{\ϵ})T_u^n+(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]---\left(7\right)$

Can be obtained grey body surface true temperature by above-mentioned (5) formula is T₀, T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-\mathrm{\ϵ})T_u^n-(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]\}}^{\frac{1}{n}}---\left(8\right)$

Step (1.5), calculate 2 true temperatures of testee difference particularly as follows:

Closely during thermometric, ignore the impact of atmospheric transmittance, make τ_a=1, the formula (7) in the most described step 1.4 becomes

$T_r=T_0{\{\mathrm{\ϵ}\[1-{\left(\frac{T_u}{T_0}\right)}^n\]+{\left(\frac{T_u}{T_0}\right)}^n\}}^{\frac{1}{n}}---\left(9\right)$

Formula (8) becomes

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[T_r^n-(1-\mathrm{\ϵ})T_u^n\]\}}^{\frac{1}{n}}---\left(10\right)$

When object surface temperature is much larger than ambient temperature, i.e. T_u/T₀＞ ＞ 0, then (9), (10) formula become

$T_r={\mathrm{\ϵ}}^{\frac{1}{n}}{T}_0---\left(11\right)$

$T_0=\frac{T_r}{\sqrt[n]{\mathrm{\ϵ}}}---\left(12\right)$

If the radiation temperature measuring testee surface any two points is T respectively_r1And T_r2, then this true temperature of 2 Difference is

$\mathrm{\Δ}T=T_{01}-T_{02}=\frac{1}{{\mathrm{\ϵ}}^{1/n}}\{{\[T_{r1}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}-{\[T_{r2}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}\}---\left(13\right)$

Step 2 particularly as follows:

Step (2.1), when in step 1 between any two points of testee surface shell temperature change reach thermal infrared imager Resolution time, use thermal infrared imager detection and recorded this change, the position of show abnormality high temperature or low temperature, aobvious Demonstrate the temperature distribution image of body surface with the form of pseudo-colours in display screen.

Step (2.2), use high-precision blackbody furnace as standard before test object surface temperature, draw out photoelectricity Switching device output signal and the relation curve of blackbody furnace temperature, the relation right and wrong between emittance and the temperature of black body emission Linear, it is calculated by thermal imaging system spectral response and Planck's law of radiation, in order to set up between amount of radiation and temperature Relation, carries out different temperatures and arranges and measure it, and carried out by accurate to measurement result and black matrix temperature value black matrix Matching, has just obtained calibration curve, by the sensor on camera lens, the infrared energy of testee is converted into the signal of telecommunication, These raw electrical signals are processed and are transformed into gray value by subsequent conditioning circuit after camera lens further again, demonstrate at display Come, through such a process, just obtained the corresponding relation between image intensity value and object temperature, and then obtained gray-scale map Picture.

Step 3, specifically, Medical Image Processing algorithm research, is being gone based on the wavelet package transforms image revising Wiener filtering Make an uproar, small echo self-adaptive spot noise based on Pulse-coupled Neural Network Model filters, based on medical image increasings such as wavelet transformations Being proposed for the rim detection new algorithm that noise is more sane on the basis of strong algorithms, it is based on Corner Detection sane to noise to design Edge detector, utilize multiple dimensioned anisotropic Gaussian direction gradient wave filter to extract edge and characteristic point structural information:

Step (3.1), set up different types of edge model, including staged, pulsed, ramp type and double ridge edge Model, and the model of characteristic point, including L-type, Y type, X-type, the model of characteristic point of star-like, end type, notch cuttype, then set up Can effectively distinguish the anisotropic Gaussian direction gradient relative to different local feature models of adjacent edge and characteristic point Expression formula, and the measure of feature degree of isolation, carry out the measurement of noise sensitivity and estimation and in geometry and gray scale The tolerance of robustness in conversion；

Step (3.2), link anisotropic Gaussian direction gradient represent；

Step (3.3), infrared image is carried out logarithmic transformation, make speckle noise be converted to additive noise, by infrared image Process through Wiener filtering, calculate the standard variance of its additive noise, in this, as wavelet threshold, then, utilize wavelet transformation pair Image carries out pretreatment, utilizes PCNN to revise wavelet coefficient accordingly in wavelet field, finally, carries out wavelet inverse transformation And exponential transform, it is thus achieved that filter the image of noise.

Test simulation:

In test experiments, the distance between testee and detector is set to 3m, and ambient temperature is respectively 13 DEG C, and 20 DEG C, about 22 DEG C, during experiment, first adjust the temperature of blackbody furnace, keep blackbody furnace temperature stabilization, record black matrix thermography, find The gray value of corresponding point on image, experiment is measured in the range of blackbody temperature 293K～337K, is recorded under the conditions of varying environment Experimental data count in table 1 to table 3 respectively:

Table 1 temperature and gray scale corresponding relation experimental data (ambient temperature: 13 DEG C)

Table 2 temperature and gray scale corresponding relation experimental data (ambient temperature: 20 DEG C)

Table 3 temperature and gray scale corresponding relation experimental data (ambient temperature: 22 DEG C)

The data that table 1～table 3 obtain are to obtain under different ambient temperatures.Between blackbody temperature and the gray scale of image There is certain corresponding relation, it was demonstrated that utilize thermal imaging system thermometric feasible.Although data have certain deviation, but the trend of curve is substantially Identical.By analyzing measurement result, find that its accuracy is affected by operating ambient temperature bigger, so utilizing thermal imaging system thermometric Temperature measuring model should be used at a temperature of specific environment.

In this method, step carries out wavelet inverse transformation and exponential transform in (3.3), it is thus achieved that filter the image of noise, selects Mean square error (MSE) and two parameters of Y-PSNR (PSNR) are in this, as weighing the standard of filtering performance, to cervical myoma Ultrasonoscopy adds the speckle noise of varying strength, by method in this paper and classical denoising method Wiener filtering (Wiener), the median filter method (PCNN-MF) of Based PC NN and the wavelet soft-threshold filter method of Based PC NN (PCNN-WD) being respectively processed, comparative result is as shown in table 4.Image wherein adds the speckle that noise variance is 0.1 make an uproar Shown in sound, filter effect Fig. 1 (a) of each method～Fig. 1 (f).

The denoising effect of each filtering method under the different noise intensity of table 4

In order to verify the effectiveness of this method, use the based on Wiener filtering little of current typical Denoising Algorithm and proposition Ripple packet transform image de-noising method carries out contrast experiment.Typical Denoising Algorithm includes Wiener filtering denoising and wavelet packet self adaptation Two kinds of methods of threshold denoising.In an experiment, select vertical phrenic lymph nodes core CT figure as testing object, as shown in Figure 2.To test figure Being 0 as being separately added into average, variance is the white Gaussian noise of 0.05, and then noise image carries out WAVELET PACKET DECOMPOSITION, selects db4 Small echo, the Decomposition order arranging small echo is 2, threshold function table use soft-threshold function, estimation use revised Wiener filtering Method.Image after processing the denoising obtained is the most as shown in Figures 3 to 6.

Test result indicate that: when noise variance is 0.01, after this algorithm denoising, the PSNR of image is more adaptive than wavelet packet The PSNR after threshold denoising is answered to exceed 8.8dB.This algorithm can not only remove additive white Gaussian noise effectively, and can be well Retain marginal information, significantly improve the visual quality of image.

In order to improve medical image quality, on the basis of analysis wavelet shift theory, the imaging according to medical image is special Levy, it is proposed that a kind of medical image enhancement method based on Wavelet Fusion technology.First, image to be reinforced is carried out multi-level Wavelet Transform Conversion process, obtains the wavelet coefficient of each frequency.After the most each coefficient of frequency being processed accordingly, carry out small echo Reconstruct and carry out contrast enhancement processing, it is thus achieved that strengthening image 1；Image to be reinforced is carried out logarithmic transformation simultaneously and contrast is drawn Stretch process, it is thus achieved that strengthen image 2.Finally enhancing image 1 and enhancing image 2 are converted into wavelet field and carry out image co-registration process, To obtain final enhancing image.

Result shows: Enhancement Method in this paper has obvious reinforced effects.This Enhancement Method can be effectively improved doctor Learn the contrast of image, strengthen edge detail information, the position of prominent focus point, reach preferable reinforced effects, for medical treatment work Author observes disease offer and becomes apparent from foundation accurately.

According to the method proposed, in the environment of MATLAB 2010a, vertical phrenic lymph nodes core CT figure is used to survey as experiment Attempting sheet, test Enhancement Method, experimental result will compare with artwork, propose the effective of Enhancement Method with checking Property.

Accurately contrasted by the infrared imaging figure of the infrared medicine image after step 2 is processed with corresponding health tissues, Integrated structure image technology, determines whether human organ, tissue etc. occur pathological changes, can accurately understand the position of focus simultaneously, Size, temperature conditions, thus reach the effect of more preferable medical pathology diagnosis.

Above description of the present invention is part case study on implementation, but the invention is not limited in above-mentioned specific embodiment party Formula.Above-mentioned detailed description of the invention is schematic, is not restrictive.The method of every employing present invention, without departing from In the case of present inventive concept and scope of the claimed protection, within all concrete expansions all belong to protection scope of the present invention.

Claims (4)

1. a tissue focus detection method based on Infrared Thermography Technology, it is characterised in that specifically according to following steps Implement:

Step 1, set up infrared measurement of temperature physical model；

Step 2, utilize physical model analysis temperature and the corresponding relation of gray scale that step 1 obtains, will be obtained by thermal infrared imager To image be converted into gray level image；

Step 3, the gray level image obtaining described step 2 process；

Step 4, the result that described step 3 obtains being analyzed with medical image, finder soma is sick early Stove.

A kind of tissue focus detection method based on Infrared Thermography Technology the most according to claim 1, its feature exists In, described step 1 is specifically implemented according to following steps:

Step (1.1), first trying to achieve radiant illumination according to infrared radiation temperature formula, formula is as follows:

E_λ=A₀d^-2[τ_aλε_λL_bλ(T₀)+τ_aλ(1-α_λ)L_bλ(T_u)+ε_aλL_bλ(T_a)] (1)

(1) in formula, α_λFor testee Surface absorption rate, ε_λFor testee slin emissivity, ε_aλFor atmospheric emission rate, τ_aλFor The spectral-transmission favtor of air, T₀For object surface temperature, T_aFor atmospheric temperature, T_uFor ambient temperature, d is measured object body surface Face is to the vertical dimension between measuring instrument, A₀d^-2For constant, A₀Effective area for target；

Step (1.2), the radiant illumination obtained according to step (1.1) try to achieve the response voltage of thermal infrared imager:

Generally infrared thermography is operated in a wavelength band the narrowest, between 8～14 μm or between 3～5 μm, then recognizes For the ε in (1) formula_λ、α_λ、τ_aλUnrelated with λ, then obtain the response voltage V of thermal infrared imager_SFor:

$V_S=A_R{A}_0{d}^{-2}\{{\mathrm{\τ}}_a\[\mathrm{\ϵ}\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_0\right)d\mathrm{\λ}+(1-a)\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_u\right)d\mathrm{\λ}\]+{\mathrm{\ϵ}}_a\underset{{\mathrm{\λ}}_1}{\overset{{\mathrm{\λ}}_2}{\∫}}{R}_{\mathrm{\λ}}{L}_{b\mathrm{\λ}}\left(T_a\right)d\mathrm{\λ}\}---\left(2\right)$

(2) in formula, A_RFor the area of thermal imaging system lens, A₀For the effective area of target, A₀d^-2For constant, make K=A_RA₀d^-2,Then (2) formula becomes

V_S=K{ τ a [ε f (T₀)+(1-a)f(T_u)+ε_af(T_a)]} (3)

Step (1.3), calculating body surface true temperature:

According to Planck blackbody radiation law, can obtain

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-a)T_u^n\]+{\mathrm{\ϵ}}_a{T}_a^n---\left(4\right)$

(4) in formula, T_rFor the radiation temperature on testee surface, T_oFor the true temperature on testee surface, T_uFor environment temperature Degree, T_aFor atmospheric temperature, α is testee Surface absorption rate, and ε is surface emissivity, ε_aFor atmospheric emission rate, τ_aFor greatly The absorbance of gas, the value of n is different with wave band difference when using thermal imaging system, and to InSb (3～5 μm) detector, n value is 8.68, To HgCdTe (6～9 μm) detector, n value is 5.33, and to HgCdTe (8～14 μm) detector, n value is 4.09；

If testee surface true temperature is T₀, then T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-a)T_u^n-\frac{{\mathrm{\ϵ}}_a}{{\mathrm{\τ}}_a}{T}_a^n\]\}}^{\frac{1}{n}}---\left(5\right)$

Step (1.4), when human body surface emissivity keep constant, human body can be seen as grey body, when testee surface meet During grey body approximation, i.e. ε=α, and if thinking air ε_α=α_a=1-τ_a, then the true temperature of ash surface is calculated as follows:

By above-mentioned (3) Shi Ke get:

V_S=K{ τ_a[εf(T₀)+(1-ε)f(T_u)]+(1-τ_a)f(T_a)} (6)

By above-mentioned (4) Shi Ke get:

$T_r^n={\mathrm{\τ}}_a\[{\mathrm{\ϵT}}_0^n+(1-\mathrm{\ϵ})T_u^n+(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]---\left(7\right)$

Can be obtained grey body surface true temperature by above-mentioned (5) formula is T₀, T₀Computing formula as follows:

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[\frac{1}{{\mathrm{\τ}}_a}{T}_r^n-(1-\mathrm{\ϵ})T_u^n-(\frac{1}{{\mathrm{\τ}}_a}-1)T_a^n\]\}}^{\frac{1}{n}}---\left(8\right)$

Step (1.5), calculate 2 true temperatures of testee difference particularly as follows:

Closely during thermometric, ignore the impact of atmospheric transmittance, make τ_a=1, the formula (7) in the most described step 1.4 becomes

$T_r=T_0{\{\mathrm{\ϵ}\[1-{\left(\frac{T_u}{T_0}\right)}^n\]+{\left(\frac{T_u}{T_0}\right)}^n\}}^{\frac{1}{n}}---\left(9\right)$

Formula (8) becomes

$T_0={\{\frac{1}{\mathrm{\ϵ}}\[T_r^n-(1-\mathrm{\ϵ})T_u^n\]\}}^{\frac{1}{n}}---\left(10\right)$

When object surface temperature is much larger than ambient temperature, i.e. T_u/T₀＞ ＞ 0, then (9), (10) formula become

$T_r={\mathrm{\ϵ}}^{\frac{1}{n}}{T}_0---\left(11\right)$

$T_0=\frac{T_r}{\sqrt[n]{\mathrm{\ϵ}}}---\left(12\right)$

If the radiation temperature measuring testee surface any two points is T respectively_r1And T_r2, then this true temperature difference of 2 is

$\mathrm{\Δ}T=T_{01}-T_{02}=\frac{1}{{\mathrm{\ϵ}}^{1/n}}\{{\[T_{r1}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}-{\[T_{r2}^n-(1-\mathrm{\ϵ})T_u^n\]}^{1/n}\}---\left(13\right).$

A kind of tissue focus detection method based on Infrared Thermography Technology the most according to claim 1, its feature exists In, described step 2 particularly as follows:

Step (2.1), when in described step 1 between any two points of testee surface shell temperature change reach thermal infrared imager Resolution time, use thermal infrared imager detection and recorded this change, the position of show abnormality high temperature or low temperature, aobvious Demonstrate the temperature distribution image of body surface with the form of pseudo-colours in display screen；

Step (2.2), use high-precision blackbody furnace as standard before test object surface temperature, draw out opto-electronic conversion Device output signal and the relation curve of blackbody furnace temperature, the relation between emittance and the temperature of black body emission is non-linear , it is calculated by thermal imaging system spectral response and Planck's law of radiation, in order to set up the relation between amount of radiation and temperature, Black matrix carries out different temperatures arrange it is measured, and accurate to measurement result and black matrix temperature value is fitted, Just obtain calibration curve, by the sensor on camera lens, the infrared energy of testee is converted into the signal of telecommunication, then warp Cross camera lens subsequent conditioning circuit below these raw electrical signals are processed further to be transformed into gray value, show at display, Through such a process, just obtain the corresponding relation between image intensity value and object temperature, and then obtained gray level image.

A kind of tissue focus detection method based on Infrared Thermography Technology the most according to claim 1, its feature exists In, described step 3, specifically, Medical Image Processing algorithm research, is being gone based on the wavelet package transforms image revising Wiener filtering Make an uproar, small echo self-adaptive spot noise based on Pulse-coupled Neural Network Model filters, based on medical image increasings such as wavelet transformations Being proposed for the rim detection new algorithm that noise is more sane on the basis of strong algorithms, it is based on Corner Detection sane to noise to design Edge detector, utilize multiple dimensioned anisotropic Gaussian direction gradient wave filter to extract edge and characteristic point structural information:

Step (3.1), set up different types of edge model, including staged, pulsed, ramp type and double ridge edges mould Type, and the model of characteristic point, including L-type, Y type, X-type, the model of characteristic point of star-like, end type, notch cuttype, then set up energy Effectively distinguish adjacent edge and the anisotropic Gaussian direction gradient table relative to different local feature models of characteristic point Reach formula, and the measure of feature degree of isolation, carry out measurement and the estimation of noise sensitivity and become at geometry and gray scale Change the tolerance of middle robustness；

Step (3.2), link anisotropic Gaussian direction gradient represent；

Step (3.3), infrared image is carried out logarithmic transformation, make speckle noise be converted to additive noise, by infrared image through dimension Receive Filtering Processing, calculate the standard variance of its additive noise, in this, as wavelet threshold, then, utilize wavelet transformation to image Carry out pretreatment, utilize PCNN in wavelet field, wavelet coefficient to be revised accordingly, finally, carry out wavelet inverse transformation and refer to Transformation of variables, it is thus achieved that filter the image of noise.