阅读说明：本技术 一种纤维材料缺陷检测方法 (Fiber material defect detection method ) 是由 付乐伟 于 2019-10-31 设计创作，主要内容包括：本发明属于纤维材料缺陷检测领域,尤其涉及一种纤维材料缺陷检测,利用对样本分层准备,利用高光谱成像仪获取样本高光谱图像时,效剔除部分噪声干扰,有效的检测出了纤维材料的缺陷问题。(The invention belongs to the field of fiber material defect detection, and particularly relates to fiber material defect detection.)

1. A fiber material defect detection method is characterized by comprising the following steps:

a, sample layering preparation, namely preprocessing the layered sample, eliminating interference information, and distinguishing useless information from useful information before further data analysis;

b, when a hyperspectral imager is used for acquiring a hyperspectral image of a sample, effectively eliminating partial noise interference, after relevant parameters of a hyperspectral imaging system are adjusted, firstly acquiring a standard all-black reference image and a standard all-white (the reflectivity is more than 99%) reference image, and performing black-and-white correction on the acquired original sample image by using a formula;

c, adjusting system parameters of the hyperspectral classifier: the speed of the electric moving platform is 0.28 cm/s, the exposure time of the camera is 10ms, and the target distance is 15 cm;

d, before collecting the image, collecting a standard plate image for sample data correction;

e, flatly placing the sample on the object carrying platform, facing the spectrum camera, collecting the whole information of the sample through the movement of the electric platform, and after the electric platform returns to the original position, performing black-and-white correction on the experimental sample image by using the standard plate image obtained in the step (3), and finishing the sample data collection for one time;

f, performing multivariate scattering correction on the spectrum, firstly selecting a window with the size of n by an S-G convolution smoothing algorithm, not extracting a data value in each window by moving the window, and calculating an average value, thereby calculating an ideal spectrum value of the spectrum data point, reducing errors and improving the signal-to-mouth ratio.

Technical Field

The invention belongs to the field of fiber material defect detection, and particularly relates to fiber material defect detection.

Background

The fiber material is a multiphase solid material with excellent performance formed by integrating two or more materials with different properties through a certain process. In practical applications, the fiber material may cause some defects in some components, which causes quality problems and even results in the rejection of the whole component, resulting in significant losses. Therefore, the nondestructive testing technology specially aiming at the hot spot of the fiber material is deeply researched, and the nondestructive testing method makes a great contribution to the nondestructive testing development of the fiber material and the application and popularization of the composite material.

Disclosure of Invention

The invention aims to provide a fiber material defect detection method

In order to achieve the purpose, the invention adopts the following technical scheme.

a, sample layering preparation, namely preprocessing the layered samples, eliminating interference information, and distinguishing useless information from useful information before further data analysis.

And b, effectively eliminating part of noise interference when the hyperspectral imager is used for acquiring a hyperspectral image of the sample. After relevant parameters of the hyperspectral imaging system are adjusted, a standard all-black reference image and a standard all-white (the reflectivity is more than 99%) reference image are obtained, and black and white correction is performed on the collected sample original image by using a formula.

c, adjusting system parameters of the hyperspectral classifier: the speed of the motorized stage was 0.28 cm/sec, the exposure time of the camera was 10ms, and the target distance was 15 cm.

And d, before the image is collected, collecting a standard plate image for sample data correction.

e, horizontally placing the sample on the object carrying platform, facing the spectrum camera, and collecting the overall information of the sample through the movement of the electric platform. And (4) after the electric platform is reset, performing black and white correction on the experimental sample image by using the standard plate image obtained in the step (3), and finishing one-time sample data acquisition.

f, performing multivariate scattering correction on the spectrum, firstly selecting a window with the size of n by an S-G convolution smoothing algorithm, not extracting a data value in each window by moving the window, and calculating an average value, thereby calculating an ideal spectrum value of the spectrum data point, reducing errors and improving the signal-to-mouth ratio.

The beneficial effects are that: the defect problem of the fiber material is effectively detected.

Detailed Description

The invention is described in detail below with reference to specific embodiments.

The specific implementation mode is as follows:

a, sample preparation, namely preprocessing a sample, eliminating interference information, and distinguishing useless information from useful information before further data analysis.

And b, effectively eliminating part of noise interference when the hyperspectral imager is used for acquiring a hyperspectral image of the sample. After relevant parameters of the hyperspectral imaging system are adjusted, a standard all-black reference image and a standard all-white (the reflectivity is more than 99%) reference image are obtained, and black and white correction is performed on the collected sample original image by using a formula.

c, adjusting system parameters of the hyperspectral classifier: the speed of the motorized stage was 0.28 cm/sec, the exposure time of the camera was 10ms, and the target distance was 15 cm.

And d, before the image is collected, collecting a standard plate image for sample data correction.

e, horizontally placing the sample on the object carrying platform, facing the spectrum camera, and collecting the overall information of the sample through the movement of the electric platform. And (4) after the electric platform is reset, performing black and white correction on the experimental sample image by using the standard plate image obtained in the step (3), and finishing one-time sample data acquisition.

f, performing multivariate scattering correction on the spectrum, firstly selecting a window with the size of n by an S-G convolution smoothing algorithm, not extracting a data value in each window by moving the window, and calculating an average value, thereby calculating an ideal spectrum value of the spectrum data point, reducing errors and improving the signal-to-mouth ratio.

Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the protection scope of the present invention, although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.