Quantitative detection of goose and duck mixed velvet by hyperspectral camera

In the textile industry, goose down and duck down have become high-quality raw materials for making high-grade thermal products because of their excellent thermal properties. However, there is a big difference in the market price between goose down and duck down. Some bad merchants often mix duck down into goose down for the pursuit of high profits, which not only damages the interests of consumers, but also disrupts the market order. Therefore, accurate and efficient quantitative detection of goose and duck mixed velvet is particularly important. In recent years, the development of hyperspectral camera technology has provided an innovative solution to this detection challenge.

一、Sample preparation: Collect a large number of pure goose down and duck down samples to ensure that their sources are reliable and representative. Use high-precision electronic scales to accurately weigh goose down and duck down according to different proportions, and configure a series of goose and duck mixed velvet samples with known mixing proportions, such as setting 5%, 10%, 15%... Samples of different proportions such as 95% duck down were mixed, and multiple repeated samples were set for each proportion to improve the accuracy and reliability of the experiment. The configured mixed wool sample is evenly laid on the special sample table to ensure uniform sample distribution without overlap and vacancy, and to ensure that the hyperspectral camera can obtain comprehensive and accurate spectral information.

二、Hyperspectral image acquisition: This paper uses a 400-1000nm hyperspectral camera, which can be used for related research FS13, the product of Hangzhou Color Spectrum Technology Co., LTD. The spectral range is 400-1000nm, the wavelength resolution is better than 2.5nm, and up to 1200 spectral channels can be reached. The acquisition speed can reach 128FPS in the full spectrum, and the maximum after band selection is 3300Hz (support multi-region band selection). Each mixed wool sample is photographed multiple times to obtain images from different angles to reduce detection errors caused by local feature differences of the sample. After each shooting, the acquired hyperspectral image data is transferred to the computer for storage in time to avoid data loss.

三、Data preprocessing: The use of professional data processing software to pre-process the hyperspectral image data collected. First of all, the radiation correction is carried out to eliminate the radiation error caused by the performance difference of the camera itself and environmental factors, so that the spectral data between different images are comparable. Geometric correction is then performed to correct the image distortion caused by the camera Angle, sample placement, etc., to ensure that the position of each pixel in the image is accurate. The image is denoised, and the noise interference in the image is removed by filtering algorithm to improve the quality and clarity of the image, so as to extract the spectral features more accurately.

四、Spectral feature extraction: Specific algorithms and software tools are used to extract the spectral features of goose down and duck down regions respectively based on the pre-processed hyperspectral images. Through the analysis and comparison of a large number of image data, it is determined that the specific wavelength range of goose down and duck down can be distinguished significantly in the visible light to near infrared spectrum. At these key wavelengths, the reflectance values of goose down and duck down are carefully measured and recorded to form their own unique spectral feature data sets. For example, after many experimental analyses, it has been found that there are obvious differences in the reflectance curves of goose down and duck down in the wavelength range of 700nm-800nm, which can be used as an important basis for identifying the two.

五、Model establishment and verification: Based on the extracted spectral characteristic data of goose down and duck down, the spectral model for quantitative analysis of goose and duck mixed down was established by using machine learning or statistical methods. Common modeling methods include support vector machine, partial least square method and so on. In the process of modeling, a part of sample data with known mixing ratio is used as a training set to train the model, so that it can learn the internal relationship between the spectral characteristics of goose down and duck down and the mixing ratio. Another part of the sample data that did not participate in the training was used as the verification set to verify the established model. The hyperspectral image data of the validation set samples were input into the model, and the predicted mixing ratio of goose down and duck down was calculated by the model, and compared with the actual known mixing ratio. The accuracy and reliability of the model are evaluated by calculating the error between the predicted value and the true value, such as the root-mean-square error and the average absolute error. According to the verification results, the model is adjusted and optimized, such as adjusting model parameters, adding or reducing feature variables, etc., to improve the performance of the model.

6. Analysis and evaluation of results: The test results of all mixed wool samples were summarized and statistically analyzed. Statistical indexes such as mean value and standard difference of test results under different mixing ratio were calculated to evaluate the stability and repeatability of the test method. The results of hyperspectral camera detection were compared with those of traditional detection methods (such as chemical analysis) to further verify the accuracy of the hyperspectral camera detection method. Through the analysis of a large number of experimental data, the error range, detection accuracy and other key performance indexes of the hyperspectral camera in quantitative detection of goose and duck mixed velvet are obtained. The experimental results show that the method can quickly and accurately detect the precise proportion of goose down and duck down in mixed velvet in a short time, and the detection error can be effectively controlled in a very small range, which fully demonstrates its high reliability and practicability.

The application of hyperspectral camera technology greatly improves the accuracy and efficiency of quantitative detection of goose and duck mixed velvet. For production enterprises, it can ensure product quality and maintain brand reputation; For the regulatory authorities, it provides strong technical support for cracking down on counterfeit and shoddy products in the market, which helps to purify the market environment and protect the legitimate rights and interests of consumers. With the continuous development and improvement of technology, it is believed that the application of hyperspectral cameras in the quantitative detection of goose and duck mixed velvet and other related fields will be more extensive and in-depth, and inject new vitality into the healthy development of the industry.