CN-121521785-B - Metal element detection method for metal material

CN121521785BCN 121521785 BCN121521785 BCN 121521785BCN-121521785-B

Abstract

The invention relates to the technical field of detection of metal material components, and discloses a metal element detection method for a metal material. The method comprises the steps of driving a probe to scan and collect an original spectrum signal flow by establishing a dynamic detection path for adjusting the spatial distribution of detection points in real time according to the material morphology. And carrying out real-time quality evaluation on the signal flow, and selectively activating a signal enhancement mechanism according to an evaluation result to generate a high-quality signal sequence. And decoupling characteristic components of each element from the sequence by using an element characteristic decoupling network, calculating a quantization index, and matching with a standard reference to identify abnormal elements. And starting a depth verification process adopting the differential excitation parameters for detecting the abnormal elements again. Finally, the initial result and the verification result are fused, an element content distribution map is generated, and a material database is updated. The method realizes self-adaptive optimization of the detection process and accurate rechecking of abnormal results, and improves the detection accuracy and reliability.

Inventors

WANG ZHANGYONG
LU HUI
CAI WENLING
Zhong Kaiqun
YANG HONG

Assignees

四川精迅产品质量检测有限公司

Dates

Publication Date: 20260508
Application Date: 20260115

Claims (8)

1. A metal element detection method for a metal material, characterized by comprising: Establishing a dynamic detection path for the metal material, wherein the dynamic detection path adjusts the spatial distribution of detection points in real time according to the material morphology; driving a detection probe to scan according to the dynamic detection path, and synchronously collecting an original spectrum signal flow returned by the detection probe; performing signal quality evaluation on the original spectrum signal flow, and activating a signal enhancement mechanism according to an evaluation result to generate an enhanced signal sequence; inputting the enhanced signal sequence into an element characteristic decoupling network to decouple characteristic components corresponding to different metal elements; calculating a quantization index of each metal element based on the decoupled feature components; Matching the quantization index with a reference range in a material standard database, and identifying abnormal elements with deviation; starting a depth verification process aiming at the abnormal element, wherein the depth verification process adopts differentiated excitation parameters to re-detect; Fusing the initial detection result with the output of the depth verification process to generate a final element content distribution map; updating a historical record in a new material standard database according to the final element content distribution map; The establishing a dynamic detection path for the metal material includes: acquiring three-dimensional surface topology data of a metal material, and extracting curvature characteristics of the three-dimensional surface topology data; Determining a high-probability defect area according to the curvature characteristics, and arranging a dense detection lattice in the high-probability defect area; generating a sparse detection lattice in the non-high probability defect area by adopting a self-adaptive meshing algorithm; performing path optimization splicing on the dense detection lattice and the sparse detection lattice to form a complete dynamic detection path; The depth verification process comprises the following steps: adjusting energy parameters of an excitation source, and exciting by adopting a step-type energy scanning mode; changing the acquisition angle of the detection probe to acquire a multi-angle spectrum signal; Independently analyzing the multi-angle spectrum signals to obtain a verification concentration value set; And calculating a deviation coefficient of the verification concentration value set and the initial quantization index, and confirming the validity of the abnormal element detection result when the deviation coefficient is in an allowable range.
2. A metallic element detecting method for metallic material as recited in claim 1, wherein said evaluating the signal quality of the raw spectral signal stream includes: Carrying out segmentation processing on the original spectrum signal flow, and calculating the signal-to-noise ratio index of each segment of signal; triggering a signal enhancement mechanism when the signal-to-noise ratio index is lower than a preset threshold; the signal enhancement mechanism adopts a multi-scale filtering algorithm to carry out noise reduction treatment on the signals; Baseline correction is carried out on the noise-reduced signal, and background interference components are eliminated; And splicing the processed signal segments into an enhanced signal sequence through a signal reconstruction algorithm.
3. A metallic element detection method for metallic material as recited in claim 2, wherein the operation of inputting the enhanced signal sequence into the elemental signature decoupling network comprises: Constructing a characteristic template library containing a plurality of metal element standard spectrums; Decomposing the enhanced signal sequence into linear combinations of characteristic templates by adopting a sparse coding algorithm; extracting a weight coefficient corresponding to each characteristic template through an orthogonal matching pursuit algorithm; judging a feature template with a weight coefficient larger than an activation threshold as an effective element feature; a feature component set is generated from the active element features.
4. A metal element detection method for a metal material according to claim 3, wherein the calculating a quantization index for each metal element comprises: carrying out integral intensity calculation on each characteristic component to obtain an original intensity value; Correcting the original intensity value by adopting an internal standard method, and eliminating the fluctuation influence of the instrument; converting the corrected intensity value into an element concentration value through a standard curve; Carrying out statistical consistency test on the concentration values detected for the same element for multiple times; The arithmetic average of the concentration values passing the consistency test is taken as the final quantization index.
5. The method for detecting a metal element for a metal material according to claim 4, wherein the generating a final element content distribution map includes: mapping the confirmed effective concentration value of the abnormal element back to the corresponding dynamic detection path coordinate; constructing an element concentration distribution curved surface of the whole material surface by adopting a Kriging interpolation algorithm; Performing contour line extraction on the element concentration distribution curved surface, and marking a concentration gradient change area; The distribution curved surfaces of various metal elements are integrated into a multi-level element content distribution map.
6. The method for detecting metallic elements as recited in claim 5, wherein said updating the material standard database according to the final element content distribution profile comprises: extracting statistical characteristic parameters in the final element content distribution map, and performing similarity matching on the statistical characteristic parameters and historical data in a material standard database; When the similarity is lower than the update threshold value, starting a database self-adaptive update mechanism, and reserving the detection data of the latest batch by adopting a sliding window algorithm; the upper and lower limits of the reference range in the material standard database are recalculated.
7. The method for detecting metal elements in metal material according to claim 6, wherein the signal enhancement mechanism performs noise reduction processing on the signal by using a multi-scale filtering algorithm, comprising: Performing wavelet transformation on the original spectrum signal flow, and decomposing the signal into wavelet coefficient sequences with multiple scales; For each scale wavelet coefficient sequence, calculating a scale dependent noise threshold, and performing soft threshold processing on the wavelet coefficients to eliminate noise components; and reconstructing the wavelet coefficient sequence after the threshold processing into a signal segment after noise reduction through wavelet inverse transformation.
8. The method for detecting metal elements of metal material according to claim 7, wherein the extracting the weight coefficient corresponding to each feature template by the orthogonal matching pursuit algorithm comprises: initializing residual errors to be an enhanced signal sequence, initializing an activated atom set to be an empty set, and initializing a weight coefficient vector to be a zero vector; iteratively performing the following steps until the norm of the residual is below a preset threshold or the maximum number of iterations is reached: calculating the inner product of the current residual error and each feature template in the feature template library, and selecting the feature template with the largest absolute value of the inner product as an activation atom of the current iteration; adding a current activation atom to the set of activation atoms; Solving the linear approximation of the activated atom set to the enhanced signal sequence by a least square method, and calculating a weight coefficient corresponding to the current activated atom set; updating the residual error to be the difference between the weighted sum of the enhanced signal sequence and the activated atom set; And outputting the finally obtained weight coefficient vector as a weight coefficient corresponding to the characteristic template.

Description

Metal element detection method for metal material Technical Field The invention relates to the technical field of detection of metal material components, in particular to a metal element detection method for a metal material. Background In the process of element detection and analysis of metal materials, spectroscopic analysis techniques are widely used due to their rapid and nondestructive characteristics. However, existing detection methods typically rely on a preset fixed path or simple meshing scanning strategy for data acquisition. The static sampling mode is difficult to adapt to the complex conditions of morphological fluctuation, uneven thickness of an oxide layer and the like possibly existing on the surface of an actual metal material, so that the areas with insufficient sampling or poor signal quality in key areas are treated equally, and the acquired original spectrum signals are unstable in overall quality and poor in consistency. The prior art generally employs a uniform, fixed preprocessing procedure for the collected spectral signals, such as processing all signals using the same filtering parameters and enhancement algorithms. The processing mode cannot distinguish the quality of the signal, unnecessary calculation and even distortion can be introduced to the signal with good quality, and the signal with low signal to noise ratio or serious interference can be insufficiently processed, so that effective characteristic information cannot be effectively extracted, and the accuracy of subsequent quantitative analysis is affected. When the primary detection finds that the element content deviates from the standard range, the conventional re-detection method usually uses the repeated measurement of the same instrument parameters for several times to average values at the original detection point or the adjacent position. The simple repeated verification cannot exclude systematic errors caused by matrix effects, inter-element spectrum interference or instantaneous instrument state fluctuation in the primary detection, so that misjudgment or missed judgment risks exist in the judgment of abnormal elements. Disclosure of Invention The present invention is directed to a method for detecting metallic elements of a metallic material, which solves the problems set forth in the background art. In order to achieve the above object, the present invention provides a metal element detection method for a metal material, the method comprising: Establishing a dynamic detection path for the metal material, wherein the dynamic detection path adjusts the spatial distribution of detection points in real time according to the material morphology; driving a detection probe to scan according to the dynamic detection path, and synchronously collecting an original spectrum signal flow returned by the detection probe; performing signal quality evaluation on the original spectrum signal flow, and activating a signal enhancement mechanism according to an evaluation result to generate an enhanced signal sequence; inputting the enhanced signal sequence into an element characteristic decoupling network to decouple characteristic components corresponding to different metal elements; calculating a quantization index of each metal element based on the decoupled feature components; Matching the quantization index with a reference range in a material standard database, and identifying abnormal elements with deviation; starting a depth verification process aiming at the abnormal element, wherein the depth verification process adopts differentiated excitation parameters to re-detect; Fusing the initial detection result with the output of the depth verification process to generate a final element content distribution map; And updating the historical record in the new material standard database according to the final element content distribution map. Preferably, the establishing a dynamic detection path for the metal material includes: acquiring three-dimensional surface topology data of a metal material, and extracting curvature characteristics of the three-dimensional surface topology data; Determining a high-probability defect area according to the curvature characteristics, and arranging a dense detection lattice in the high-probability defect area; generating a sparse detection lattice in the non-high probability defect area by adopting a self-adaptive meshing algorithm; And performing path optimization splicing on the dense detection lattice and the sparse detection lattice to form a complete dynamic detection path. Preferably, the signal quality evaluation of the original spectrum signal stream includes: Carrying out segmentation processing on the original spectrum signal flow, and calculating the signal-to-noise ratio index of each segment of signal; triggering a signal enhancement mechanism when the signal-to-noise ratio index is lower than a preset threshold; the signal enhancement mechanism adopts a multi-scale filtering a