A quantitative detection method for meat quality based on the fusion of thermal imaging and visual imaging during time-series temperature change

Abstract

A method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process, and in particular, a method for fully utilizing the region of interest of a thermal image and a CCD image during a time-series temperature change process for fusion and then using a quantitative detection model to detect meat quality, including first establishing different meat quality quantitative models, and then using the model to perform quantitative detection of meat quality indicators. The present invention integrates the temperature information of the thermal image during the time-series temperature change of the sample and the color information of the CCD image, and can quickly and accurately detect the content of different quality indicators of meat, and is simple to operate, economical and convenient, and has a high market promotion and application value.

Description

Technical Field

[0001] The present invention belongs to the technical field of rapid meat quality detection, and in particular relates to a meat quality quantitative detection method that integrates thermal imaging and visual imaging in a time-series temperature change process. The detection method can be used for content detection of different quality indicators of meat. Background Art

[0002] With the improvement of living standards and changes in dietary structure, people have put forward higher requirements for meat quality and safety. In recent years, meat production and consumption have gradually increased around the world. But at the same time, high profits have also led to frequent quality problems in the meat supply chain. Different indicators can evaluate the different qualities of meat. Among them, detecting the adulteration content in adulterated meat can protect the economic interests of consumers. TVB-N content is the main chemical information for judging the freshness of meat. The nutrient content (protein, fat, inorganic matter, vitamins, etc.) is the reference for the edible value of meat. Therefore, it is necessary to test different quality indicators of meat and determine their content. Among the existing technical means of meat quality detection, DNA technology has high requirements for sample DNA purity, complex preparation, and the required instruments are expensive; spectral technology has poor sensitivity and is difficult to model samples; electronic tongue technology experimental results are not highly repeatable and have low identification accuracy; although thermal imaging and CCD technologies are simple to operate and have strong adaptability, as a single detection method, the information provided can often only describe one aspect of meat quality and cannot fully describe the overall information of meat quality; although the fusion of thermal imaging and visual imaging can reflect the temperature and color information of meat, the research has mostly focused on constant temperature conditions, and the meat characteristics reflected are relatively few, which has certain limitations. Summary of the invention

[0003] The purpose of the present invention is to provide a meat quality quantitative detection method based on the fusion of thermal imaging and visual imaging in a time-series temperature change process. Based on the surface temperature distribution differences and color change characteristics of meat of different qualities in the time-series temperature change process, an acquisition system is used to collect infrared thermal imaging videos and CCD images of different samples in the time-series temperature change process. At the same time, a meat quality quantitative model is established based on the fusion data of thermal images and CCD images combined with a convolutional neural network to detect meat quality indicators, aiming to solve the problems of complex operation of existing meat quality detection technology, expensive instruments, low accuracy of traditional recognition models, and single information acquisition.

[0004] The technical solution adopted by the present invention is as follows:

[0005] A meat quality quantitative detection method based on the fusion of thermal imaging and visual imaging in a time-series temperature change process, characterized in that a modeling sample is prepared, thermal imaging and visual imaging data in a time-series temperature change process are collected and fused, and then a meat quality quantitative detection model is established, and then the model is used to quantitatively detect the quality of meat samples, including the following steps:

[0006] Step 1: Prepare representative meat samples with consistent height, quality, and density and different quality indicators;

[0007] Step 2: Place the sample in a temperature change device for sequential heating or cooling, and simultaneously collect thermal video and visual images of the sample during the sequential temperature change process through an infrared thermal imaging acquisition system and a visual imaging acquisition system;

[0008] Step 3: Extract sample thermal images from the thermal imaging video during the time-series temperature change process, wherein a single or multiple thermal images and a corresponding time-series CCD image are extracted for each sample;

[0009] Step 4: Remove the sample image background of the thermal image and CCD image respectively, extract the centroid of the sample image, and select the region of interest with this as the center; adjust the thermal image and CCD image under the region of interest to a uniform size and fuse them;

[0010] Step 5: Use the sample fused image data as input data to divide it into a training set and a test set; use the divided sample data set to train and test the model, select the best model, and determine the various parameters of the model;

[0011] Step 6: If the effect of the constructed quantitative model meets the requirements, the model is feasible; otherwise, expand the sample set and repeat steps 1 to 5 until the requirements are met;

[0012] The specific steps of using the quantitative detection model to detect meat quality are as follows:

[0013] Step A, preparing representative meat samples with different quality indicators of uniform height, quality and density;

[0014] Step B, placing the sample in a temperature change device for sequential heating or cooling, and simultaneously collecting thermal video and visual images of the sample during the sequential temperature change process through an infrared thermal imaging acquisition system and a visual imaging acquisition system;

[0015] Step C, extracting sample thermal images from the thermal imaging video during the time-series temperature change process, wherein a single or multiple thermal images and a corresponding time-series CCD image are extracted for each sample;

[0016] Step D, removing the sample image background of the thermal image and the CCD image respectively, extracting the centroid of the sample image, and selecting the region of interest with this as the center; adjusting the thermal image and the CCD image under the region of interest to a uniform size and fusing them;

[0017] Step E: input the fused image data into the established quantitative detection model, determine and output the quantitative detection result of the sample by directly judging a single sample or comprehensively judging multiple samples.

[0018] In the above-mentioned quantitative testing of different quality indicators of meat, meat includes mutton, beef, pork, duck and other edible meats; the quality indicator testing contents are: adulterated meat content testing, nutrient content testing, TVB-N content testing, moisture content testing and microbial testing.

[0019] In the above step 1 and step A, when preparing the sample, representative meat samples with different meat qualities are prepared with uniform height, quality and density.

[0020] In the above step 2 and step B, the sample is placed in a temperature changing device for sequential heating or cooling. The sample can be heated or cooled at a constant temperature in a variety of ways, and the initial temperature of the sample remains consistent.

[0021] In the above step three and step C, when extracting the sample thermal image from the thermal imaging video, there are multiple optional extraction methods, such as: equal time intervals, equal temperature intervals, equal time intervals within a certain temperature range, and equal temperature intervals within a certain time period. The present invention prefers the first method, which continuously collects thermal images within a fixed time period. The less data in this stage contains the sample thermal images under a larger range of temperatures; when the CCD camera collects the visual image of the sample during the time-series temperature change process, a continuous shooting mode is adopted, and the image acquisition interval is consistent with the thermal imaging.

[0022] In the above step 5 and step E, different fusion methods can be used to fuse the thermal image with the CCD image, such as image stitching, data addition and channel superposition.

[0023] In the above step five, the models involved include feedforward neural networks, convolutional neural networks, and recurrent neural networks, and preferably the convolutional neural network is used to establish the model; the quantitative model classifiers involved in the present invention include Regression, SVR, and PLSR, and preferably Regression is used to detect the content of different quality indicators of meat.

[0024] In the above step 2 and step B, the detection system used by the method of the present invention includes a temperature variable unit, a lighting unit, a thermal image acquisition unit, a visual imaging acquisition unit, a dark box, a sample chamber and a data display and analysis unit; the temperature variable unit is a temperature-adjustable heating or cooling device, located inside the dark box, and is used to provide a stable temperature source for the sample; the lighting unit is a ring light source with adjustable light intensity, which is used to provide stable lighting during sample collection; the thermal image acquisition unit includes a thermal image acquisition device, a bracket and a communication module; the visual imaging acquisition unit includes an image acquisition device, a bracket and a communication module; the dark box is composed of a bracket and a thick black woven cloth, which is a closed light-proof box to prevent external light interference; the sample chamber is a black light-proof container to remove light reflection interference; the data display and analysis unit includes a PC and a software interface, wherein the software interface can realize the functions of real-time display, storage and discrimination analysis of thermal image video and CCD image.

[0025] Compared with the prior art, the present invention has the following advantages:

[0026] First, the detection method of the present invention has the advantages of being simple to operate, economical, and easy to promote.

[0027] Second, the method of the present invention uses a time-series temperature-changing device to process meat samples representative of meat quality, which can quickly obtain thermal images and CCD images of the samples at different temperatures, reflecting more meat characteristics.

[0028] Third, the fusion of thermal image and CCD image feature information in the method of the present invention can accurately quantitatively detect meat quality, thereby obtaining more comprehensive information and improving detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Schematic diagram of the device structure involved in the invention embodiment 1 and embodiment 2

[0030] Figure 2 The specific flow chart of the method of the present invention involved in Example 2 of the present invention

[0031] Figure 3 The sample thermal image and CCD image fusion process diagram involved in Example 2 of the present invention

[0032] Figure 4 This is a diagram of the convolutional neural network detection model architecture involved in Example 2 of the present invention

[0033] Figure 5 This is a distribution diagram of the predicted values ​​of the validation set of the fusion image quantitative detection model involved in Example 2 of the present invention.

[0034] The symbols in the accompanying drawings are explained as follows: 1: dark box; 2: visual imaging acquisition unit; 3: thermal imaging acquisition unit; 4: lighting unit; 5: sample chamber; 6: temperature change unit; 7: data display and analysis unit. DETAILED DESCRIPTION

[0035] In order to make the purpose and advantages of the present invention more clear, the present invention is further described below in conjunction with specific embodiments.

[0036] Example 1: A thermal video and CCD image acquisition system structure and use of a method for quantitatively detecting adulterated content in adulterated mutton based on the fusion of thermal imaging and visual imaging

[0037] Part a: A thermal video and CCD image acquisition system structure for a quantitative detection method of adulterated mutton based on the fusion of thermal imaging and visual imaging

[0038] The detection system includes a temperature changing unit, an illumination unit, a thermal image acquisition unit, a visual imaging acquisition unit, a dark box, a sample chamber, and a data display and analysis unit, such as Figure 1 As shown; the temperature variable unit is a temperature-adjustable heating or cooling device, which is located inside the dark box and is used to provide a stable temperature source for the sample; the lighting unit is a ring light source with adjustable light intensity, which is used to provide stable lighting for sample collection; the thermal image acquisition unit includes a thermal image acquisition device, a bracket and a communication module; the visual imaging acquisition unit includes an image acquisition device, a bracket and a communication module; the dark box is composed of a bracket and a thick black woven cloth, which is a closed and light-proof box to prevent external light interference; the sample chamber is a black light-proof container to remove light reflection interference; the data display and analysis unit includes a PC and a software interface, wherein the software interface can realize the real-time display, storage and discrimination analysis of thermal image video and CCD image.

[0039] Part b: A method for using a thermal video and CCD image online acquisition system for quantitative detection of adulteration content in adulterated mutton based on the fusion of thermal imaging and visual imaging.

[0040] b1. First, fill the pre-prepared constant temperature water bath with water, put it in the dark box, then turn on the constant temperature heating water bath and set its heating temperature to 70 °C .

[0041] b2. Fix the infrared radiation detection lens to the connector on the top of the bracket with bolts and adjust the lens to be perpendicular to the ground.

[0042] b3. Place the bracket together with the infrared detection and CCD image lens into the dark box, and move the bracket so that the lens is suspended vertically over the sample detection area on the water bath.

[0043] b4. Turn on the computer, open the thermal video and CCD image online acquisition software respectively, connect the infrared radiation detection lens and the CCD image lens, and identify whether the lens and computer are successfully connected through the software interface.

[0044] b5. Adjust the height of the bracket so that the height between the infrared radiation detection lens and the upper surface of the sample culture dish is 50 cm ; Adjust the height of the CCD image lens so that it is 60 degrees above the upper surface of the sample culture dish. cm .

[0045] b6. Wait until the temperature of the constant temperature water bath reaches 70 °C , and when it remains stable, place the sample to be tested in the center of the sample testing area of ​​the water bath.

[0046] b7. Remove the lens cover and fine-tune the lens so that the sample can be clearly and completely displayed in the real-time acquisition window of the software interface.

[0047] b8. Click the Start Acquisition button, and the software interface will be able to display the thermal video and CCD image of the collected sample in real time and save them.

[0048] Example 2: A quantitative detection method for adulterated mutton based on the fusion of thermal imaging and visual imaging

[0049] The specific flow chart of this embodiment is as follows Figure 2 Shown

[0050] Part C: Establishment of a quantitative detection model for the adulteration of pork and duck meat in adulterated mutton

[0051] c1. Sample preparation

[0052] The experimental materials were mutton, pork and duck meat. The fat, fascia and epidermis were removed. Pure mutton, pork and duck meat samples were prepared first. According to different adulteration ratios (10%, 20%, 30%, 40%, 50%), the corresponding weights of mutton, pork and duck meat were weighed using an electronic scale and mixed and then put into a meat grinder for 30 min. s , mince the meat into evenly granulated pieces, and then add about 30 g Put the minced meat into a 6 cm , and compacted in a round petri dish with a flat surface to make adulterated samples with uniform density. 13 samples were prepared for each adulterated ratio of each adulterated sample, and 13 samples of pure mutton, pure pork and pure duck were prepared, for a total of 169 samples.

[0053] c2. Thermal video and CCD image acquisition of samples

[0054] The detection system consists of a constant temperature heating water bath device, a bracket, an infrared radiation detection lens, a CCD image acquisition lens, a dark box, a computer and data display and analysis software.

[0055] Before data collection, turn on the constant temperature water bath and set its temperature to 70 °C When collecting adulterated mutton samples, the thermal imager (YOSEEN, China) used a spectral range of 7.5~14 μm , with an accuracy of ±2 °C , the data collection frequency is 50 Hz The image resolution is 384*288, the temperature color palette format is IronBow, and the material emissivity is set to 0.97; the CCD camera used is MV EM200M (Dimensional Image), with an optical size of 1 / 1.8, a resolution of 1600*1200, and a frame rate of 40 fps The light source used is AFT-RL150105 (Vision Image), with a rated power of 17.5 W .

[0056] Place the sample and the culture dish in a water bath and heat continuously for 10 min , a thermal imager and a CCD camera are used to collect thermal video and visible light images of the sample during the heating process, and they are displayed and saved in real time on the PC software interface.

[0057] c3. Acquisition of sample thermal image and CCD image area of ​​interest

[0058] The sample data of a single sample is a segment of 10 min Thermal imaging video with temperature data and its CCD image. First, extract single or multiple thermal images from the thermal imaging video during the sequential heating process at equal time. Use the Otsu threshold method to segment and remove the sample background, extract the edge contour of the sample after removing the background, and extract the contour centroid. Select an area of ​​85*85 with the sample center as the region of interest of the sample thermal image and its corresponding sequential CCD image. Adjust the pixels of the ROI area of ​​the sample image, use the bicubic interpolation method to resize the image, and adjust the two ROI images to the same size.

[0059] c4. Fusion of sample thermal image and CCD image area of ​​interest

[0060] The interpolated sample thermal image and the image data of the region of interest of the CCD image are fused in the form of image stitching to obtain a fused image. The extraction and fusion process of the sample thermal image and the region of interest of the CCD image is as follows: Figure 3 shown.

[0061] c5. Quantitative model establishment

[0062] The sample set is 101400 fusion images of thermal images and CCD images after sample interpolation. Firstly, the adulterated mutton samples mixed with pork and the adulterated mutton samples mixed with duck were divided into seven categories according to different adulteration ratios (0%, 10%, 20%, 30%, 40%, 50%, 100%), and then the sample data set was divided into training set and test set in a ratio of 3:1 for training and testing of quantitative models.

[0063] The ResNet-18 convolutional neural network was selected, and by comparing the learning rate and Mini-batch of the model, the optimal learning rate was determined to be 0.001 and the Mini-batch was 256. Regression was used as the mapping function, and the size of the fully connected layer was adjusted to 1*1*1. Different regression functions were established between the sample feature vector and the sample adulteration ratio, and then the regression prediction value with the smallest loss function value was selected as the final prediction value of the model. The calculation formula of the loss function is shown in (1):

[0064] (1)

[0065] in, is the regression prediction value of the model, is the actual value of the sample.

[0066] Based on the regression function combined with the optimal neural network and optimal parameters, the quantitative model of adulterated mutton based on the fusion image of pork content and duck content was established respectively. Figure 4 As shown. By calculating the correlation coefficient of the model ( ) and the root mean square error ( RMSE ) to evaluate the model effect. Correlation coefficient ( ) and the root mean square error ( RMSE ) is calculated as shown in (2) and (3):

[0067] (2)

[0068] (3)

[0069] in, represents the predicted value, represents the actual value, represents the mean of the actual values, A discrete value representing a real value. is the correlation between the model's predicted value and the actual value, The larger the value, the stronger the correlation between the two. RMSE Represents the deviation between the model's predicted value and the actual value. RMSE The smaller it is, the smaller the prediction error of the model is.

[0070] The results show that the test set of the pork content detection model established by fusion image is The value is 0.9841, and the RMSE is 0.0187; the test set of the detection model of duck meat in adulterated mutton The value is 0.9867 and the RMSE is 0.0172, which meets the requirements for quantitative detection of adulterated mutton.

[0071] d. Detection of pork and duck content in adulterated mutton

[0072] d1. Main testing process

[0073] The established quantitative model was used to detect the content of adulterated mutton. First, pure mutton, pure pork and pure duck samples were prepared, and adulterated samples with uniform density were made according to different adulteration ratios (10%, 20%, 30%, 40%, 50%). Seven samples were prepared for each adulteration ratio of each adulterated sample, and seven pure mutton, pure pork and pure duck samples were prepared, totaling 91 samples; then, the samples were placed in a temperature change device for time-sequential heating or cooling, and the thermal video and visual images of the samples in the time-sequential temperature change process were collected simultaneously through the infrared thermal imaging acquisition system and the visual imaging acquisition system, and a single or multiple thermal images of each sample and the corresponding CCD images of the time sequence were extracted; the sample image regions of interest of the thermal images and CCD images were extracted and fused respectively; finally, the fused image data was input into the established quantitative detection model, and the quantitative detection results of the samples were determined and output by direct discrimination of a single sample or comprehensive discrimination of multiple samples.

[0074] d2. Analysis of test results

[0075] The established quantitative detection model was used to quantitatively detect the pork content in adulterated mutton samples mixed with pork and the duck meat content in adulterated mutton samples mixed with duck meat. and RMSE were 0.9717 and 0.0238 respectively, and the duck meat content detection model and RMSE are 0.9616 and 0.0277 respectively. The distribution of its predicted values ​​is shown in Figure 5 shown.

[0076] The correlation coefficients of the detection models established based on thermal images, CCD images and fusion images are ( ) and root mean square error (RMSE) are shown in Table 1.

[0077] Table 1 Quantitative detection models based on different sample data sets

[0078]

[0079] The results show that the detection effect of the model established by single image data is improved compared with that of the model established by fused image data.

[0080] The present invention, through two embodiments, respectively describes the operation flow of a method for detecting adulterated content in adulterated mutton based on thermal imaging and CCD image fusion from the perspectives of the structure and use of the image acquisition system, the acquisition and fusion of the region of interest of the thermal image and CCD image, the establishment of a quantitative model, and the detection of adulterated mutton. The present invention uses a common thermal imager and a CCD camera to build a mutton quality quantitative detection platform, which can simultaneously obtain the temperature and color information of the sample during the time-series temperature change process, and is simple to operate, economical and easy to promote; the present invention uses a time-series temperature change device to process meat samples representative of mutton quality, and can quickly obtain thermal images and CCD images of the sample at different temperatures, reflecting more meat characteristics; the present invention integrates the characteristic information of the thermal image and the CCD image to accurately detect the mutton quality quantitatively, thereby obtaining more comprehensive information and improving the detection accuracy.

[0081] The detection of other food quality indicators based on the fusion technology of thermal imaging and visual imaging in the time-series temperature change process can be operated with reference to the detection method and detection process proposed in the present invention.

[0082] The above implementation modes are only used to illustrate the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions and improvements made on the basis of the technical essence of the present invention within the spirit and principles of the present invention shall be included in the scope of the present invention. The patent protection scope of the present invention is defined by the claims.

Claims

1. A method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process, characterized in that: The modeling samples are prepared, the thermal imaging and visual imaging data in the time-series temperature change process are collected and integrated, and then a meat quality quantitative detection model is established, and then the model is used to quantitatively detect the quality of meat samples, including the following steps: Step 1: Prepare representative meat samples with consistent height, quality, and density and different quality indicators; Step 2: Place the sample in a temperature change device for sequential heating or cooling, and simultaneously collect thermal video and visual images of the sample during the sequential temperature change process through an infrared thermal imaging acquisition system and a visual imaging acquisition system; Step 3: Extract sample thermal images from the thermal imaging video during the time-series temperature change process, wherein a single or multiple thermal images and a corresponding time-series CCD image are extracted for each sample; Step 4: Remove the sample image background of the thermal image and CCD image respectively, extract the centroid of the sample image, and select the region of interest with this as the center; adjust the thermal image and CCD image under the region of interest to a uniform size and fuse them; Step 5: Use the sample fused image data as input data to divide it into a training set and a test set; use the divided sample data set to train and test the model, select the best model, and determine the various parameters of the model; Step 6: If the effect of the constructed quantitative detection model meets the requirements, the model is feasible; otherwise, expand the sample set and repeat steps 1 to 5 until the requirements are met; Select ResNet-18 convolutional neural network as the quantitative detection model, with a learning rate of 0.001, a mini-batch of 256, regression as the mapping function, and a fully connected layer size of 1*1*1; The specific steps of using the quantitative detection model to detect meat quality are as follows: Step A, preparing representative meat samples with consistent height, quality, and density and different quality indicators; Step B, placing the sample in a temperature change device for sequential heating or cooling, and simultaneously collecting thermal video and visual images of the sample during the sequential temperature change process through an infrared thermal imaging acquisition system and a visual imaging acquisition system; Step C, extracting sample thermal images from the thermal imaging video during the time-series temperature change process, wherein a single or multiple thermal images and a corresponding time-series CCD image are extracted for each sample; Step D, removing the sample image background of the thermal image and the CCD image respectively, extracting the centroid of the sample image, and selecting the region of interest with this as the center; adjusting the thermal image and the CCD image under the region of interest to a uniform size and fusing them; Step E: input the fused image data into the established quantitative detection model, determine and output the quantitative detection result of the sample by directly judging a single sample or comprehensively judging multiple samples.

2. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: In the quantitative detection of different quality indicators of meat, meat includes mutton, beef, pork and duck; the quality indicator detection contents are: adulterated meat content detection, nutrient content detection, TVB-N content detection, moisture content detection and microbial detection.

3. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: In the step 1 and step A, when preparing the sample, representative meat samples with different meat qualities are prepared with uniform height, quality and density.

4. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: In step 2 and step B, the sample is placed in a temperature changing device for sequential heating or cooling, and the sample is heated or cooled at a constant temperature, and the initial temperature of the sample remains consistent.

5. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: In step 3 and step C, when extracting the sample thermal image from the thermal imaging video, the equal time interval method is selected to continuously collect thermal images within a fixed time period. The less data in this stage includes the sample thermal images under a larger temperature range. When the CCD camera collects the visual image of the sample during the time-series temperature change process, the continuous shooting mode is adopted, and the image collection interval is consistent with the thermal imaging.

6. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: In step 5 and step E, when fusing the thermal image with the CCD image, at least one of image stitching, data addition and channel superposition is selected for fusion.

7. The method for quantitatively detecting meat quality based on the fusion of thermal imaging and visual imaging during a time-series temperature change process according to claim 1, characterized in that: The detection system used in the method includes a temperature-changing unit, a lighting unit, a thermal image acquisition unit, a visual imaging acquisition unit, a dark box, a sample chamber and a data display and analysis unit; the temperature-changing unit is a temperature-adjustable heating or cooling device, which is located inside the dark box and is used to provide a stable temperature source for the sample; the lighting unit is a ring light source with adjustable light intensity, which is used to provide stable lighting when collecting samples; the thermal image acquisition unit includes a thermal image acquisition device, a bracket and a communication module; the visual imaging acquisition unit includes an image acquisition device, a bracket and a communication module; the dark box is composed of a bracket and a thick black woven cloth, which is a closed light-proof box to prevent external light interference; the sample chamber is a black light-proof container to remove light reflection interference; the data display and analysis unit includes a PC and a software interface, wherein the software interface can realize the real-time display, storage and discrimination analysis functions of thermal image video and CCD image.

Images