CN-122023906-A - Ice aggregate grading intelligent detection method based on image recognition

CN122023906ACN 122023906 ACN122023906 ACN 122023906ACN-122023906-A

Abstract

The invention relates to the technical field of image recognition and discloses an intelligent ice aggregate grading detection method based on image recognition, which comprises the steps of collecting an ultraviolet channel image and a structured light reflection image of an ice aggregate detection area; the method comprises the steps of carrying out signal extraction on an ultraviolet channel image to obtain a TCP signal characteristic image, carrying out stripe frequency analysis on a structured light reflection image to obtain a PDR stripe frequency distribution diagram, carrying out pixel-by-pixel weighting fusion calculation on the TCP signal characteristic image and the PDR stripe frequency distribution diagram to obtain a coupling strength distribution diagram, carrying out binarization judgment on an ice film region of an ice aggregate detection region according to statistical parameters of the coupling strength distribution diagram to obtain an ice film mask, generating an aggregate candidate region according to the ultraviolet channel image and the structured light reflection image, and calculating actual size parameters of ice aggregates in the ice aggregate detection region based on the aggregate candidate region and the ice film mask.

Inventors

SUN PENG
LIN YANYU
TIAN BO
Zeng Yinggen
JIANG WEIYE
YANG SHUANG
ZHANG HAN
LI SILI

Assignees

清华大学苏州汽车研究院（吴江）
中国极地研究中心(中国极地研究所)
交通运输部公路科学研究所

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. An intelligent detection method for ice aggregate grading based on image recognition is characterized by comprising the following steps: s1, acquiring an ultraviolet channel image and a structured light reflection image of an ice aggregate detection area; s2, carrying out signal extraction on the ultraviolet channel image to obtain a TCP signal characteristic image, and carrying out fringe frequency analysis on the structured light reflection image to obtain a PDR fringe frequency distribution diagram; S3, carrying out pixel-by-pixel weighting fusion calculation on the TCP signal characteristic image and the PDR stripe frequency distribution map to obtain a coupling strength distribution map; s4, performing binarization judgment on the ice film area of the ice aggregate detection area according to the statistical parameter of the coupling strength distribution diagram to obtain an ice film mask; s5, generating an aggregate candidate region according to the ultraviolet channel image and the structured light reflection image, and calculating actual size parameters of the ice aggregate in the ice aggregate detection region based on the aggregate candidate region and the ice film mask; s6, carrying out frozen water correction coefficient conversion on the average coupling strength of the coupling strength distribution diagram to obtain the frozen water content of the frozen aggregate detection area; And S7, judging whether the frozen aggregate is qualified or not according to the frozen water content and the actual size parameter.
2. The intelligent detection method for ice aggregate grading based on image recognition according to claim 1, wherein the steps of collecting an ultraviolet channel image and a structured light reflection image of an ice aggregate detection area comprise: responsive to the detection signal At moment, exciting TCP substances of the ice aggregates in the ice aggregate detection area by an ultraviolet pulse light source to generate luminescence radiation, and collecting an ultraviolet channel image of the ice aggregate detection area; At the position of At the moment of ms, stripe structure light is projected to the ice aggregate through a stripe structure light projector, and a structure light reflection image of an ice aggregate detection area is acquired.
3. The intelligent detection method for ice aggregate grading based on image recognition according to claim 1, wherein the steps of extracting the signal from the ultraviolet channel image to obtain a characteristic image of the TCP signal, and performing fringe frequency analysis on the structural light reflection image to obtain a PDR fringe frequency distribution map comprise: Collecting a background image of the aggregate-free detection area, and performing pixel-by-pixel differential operation on the ultraviolet channel image and the background image to obtain a TCP original differential image; noise cancellation processing is carried out on the TCP original differential image, so that a TCP signal characteristic image is obtained; Performing two-dimensional Fourier transform on the structure light reflection image to obtain a frequency domain amplitude spectrum of the structure light reflection image; searching energy peak coordinate point of frequency domain amplitude spectrum through peak detection algorithm Calculating the frequency value of the energy peak coordinate point; To be used for As the center, radius Defining a frequency domain region, extracting frequency components of the frequency domain region, and performing inverse Fourier transform on the frequency components to obtain a reconstructed stripe pattern; calculating an absolute phase value of each pixel point in the reconstructed fringe pattern by using a spatial phase analysis algorithm; substituting the absolute phase value into a calibration formula, and calculating the PDR stripe frequency value of each pixel point; And generating a PDR fringe frequency distribution diagram according to the PDR fringe frequency value.
4. The intelligent detection method for ice aggregate gradation based on image recognition according to claim 1, wherein the pixel-by-pixel weighted fusion calculation is performed on the characteristic image of the TCP signal and the PDR fringe frequency distribution map to obtain a coupling strength distribution map, comprising: The method comprises the steps of spatially registering a TCP signal characteristic image and a PDR stripe frequency distribution map to obtain a same-position pixel, and extracting a TCP cold light intensity value and a PDR stripe frequency value of the same-position pixel; weighting calculation is carried out on the TCP cold light intensity value and the PDR stripe frequency value to obtain a coupling intensity value; and generating a coupling strength distribution diagram according to the coupling strength values.
5. The intelligent detection method for ice aggregate grading based on image recognition according to claim 1, wherein the binarization judgment is performed on the ice film area of the ice aggregate detection area according to the statistical parameter of the coupling intensity distribution diagram to obtain the ice film mask, comprising: calculating an ice film judging threshold according to the statistical parameters; if the coupling strength value of the coupling strength distribution diagram at the pixel point is larger than the ice film judging threshold value, marking the pixel point as an ice film area and assigning 1 to the pixel point, otherwise, marking the pixel point as a non-ice film area and assigning 0 to the pixel point; And generating an ice film mask according to each pixel point of the coupling intensity distribution diagram.
6. The intelligent detection method for ice aggregate grading based on image recognition according to claim 1, wherein the generation of the aggregate candidate region according to the ultraviolet channel image and the structured light reflection image comprises the following steps: performing gray correction and Gaussian filtering noise reduction on the ultraviolet channel image and the structured light reflection image; and extracting an aggregate pre-screening area through edge detection, and screening an aggregate candidate area from the aggregate pre-screening area by using connected domain analysis.
7. The intelligent detection method for the grading of the frozen aggregate based on the image recognition according to claim 1, wherein the calculation of the actual size parameter of the frozen aggregate in the frozen aggregate detection area based on the aggregate candidate area and the frozen film mask comprises the following steps: Carrying out morphological treatment on the ice film mask; under the constraint of the aggregate candidate region, contour extraction is carried out on the ice film mask subjected to morphological treatment, so that a real aggregate contour image is obtained; And calculating the actual size parameters of the ice aggregate according to the actual aggregate contour image, wherein the actual size parameters comprise an actual perimeter, an actual length, an actual width, an actual area, an actual length-diameter ratio and an actual shape factor.
8. The intelligent detection method for ice aggregate gradation based on image recognition as claimed in claim 1, wherein the step of performing the conversion of the frozen water correction coefficient on the average coupling strength of the coupling strength distribution map to obtain the frozen water content of the ice aggregate detection area comprises the steps of: calculating the average coupling strength of the coupling strength distribution map; substituting the average coupling strength into a correction formula to obtain the frozen water content.
9. The intelligent detection method for ice aggregate grading based on image recognition as claimed in claim 8, wherein the calculation formula of the correction formula is as follows: In the formula, Is the content of the frozen water and the frozen water, Is the correction coefficient of the frozen water, Is the average coupling strength.
10. The intelligent detection method for grading of the frozen aggregate based on image recognition as recited in claim 7, wherein the step of judging whether the frozen aggregate is qualified according to the frozen water content and the actual size parameter comprises the following steps: grouping the actual size parameters of all the ice aggregates according to the particle size intervals to obtain a plurality of particle size distribution groups; Counting the quantity of the aggregate in the group of the particle size distribution group and the total quantity of the aggregate in the aggregate detection area, and calculating the ratio of the quantity of the aggregate in the group to the total quantity of the aggregate to obtain the ratio of the quantity of the aggregate in the particle size distribution group; Calculating the total in-group area of all the ice aggregates in the particle size distribution group according to the actual area, and calculating the ratio of the total in-group area to the total ice aggregate area of the ice aggregate detection area to obtain the ice aggregate area occupation ratio of the particle size distribution group; Determining the shape class of the ice aggregate according to the actual length-diameter ratio and the actual shape factor for each particle size distribution group, counting the quantity proportion of the ice aggregate in the class under the shape class, and calculating the ratio of the quantity proportion of the ice aggregate in the class to the quantity of the ice aggregate in the group to obtain the shape class proportion of the ice aggregate in the particle size distribution group; judging whether the frozen aggregate is qualified or not according to the aggregate quantity proportion, the frozen aggregate area proportion, the frozen aggregate shape class proportion and the frozen water content.

Description

Ice aggregate grading intelligent detection method based on image recognition Technical Field The invention relates to the technical field of image recognition, in particular to an intelligent ice aggregate grading detection method based on image recognition. Background In high-cold or seasonal low-temperature engineering sites, such as Gao Tiezhi beam fields, airport runway restoration and concrete prefabrication workshops of polar scientific investigation stations, crushed stone aggregate is often stored in an environment below 0 ℃ or rapidly cooled after being sprayed manually, and ice films and frozen water with uneven thickness are inevitably formed on the surfaces. When the ice aggregates are conveyed to a mixing system, the grain size grading, shape parameters and water-containing state of the ice aggregates are dynamically changed along with the thickness of the ice film. The existing detection scheme mainly relies on manual screening, weighing or single optical sensor sampling, so that the real aggregate contour cannot be identified through a semitransparent ice film, the instant quantification capability of frozen water content is lacking, and on a high-speed production line, reliable data support is difficult to provide for real-time proportioning adjustment in the hysteresis estimation type detection, so that quality hidden troubles such as slump loss control, early strength fluctuation and surface bleeding of a mixture are extremely easy to cause. Disclosure of Invention The invention provides an intelligent ice aggregate grading detection method based on image recognition, which mainly aims to solve the problem that in the prior art, hysteresis estimation type detection is difficult to provide reliable data support for real-time proportioning adjustment. In order to achieve the above purpose, the invention provides an intelligent ice aggregate grading detection method based on image recognition, which comprises the following steps: s1, acquiring an ultraviolet channel image and a structured light reflection image of an ice aggregate detection area; s2, carrying out signal extraction on the ultraviolet channel image to obtain a TCP signal characteristic image, and carrying out fringe frequency analysis on the structured light reflection image to obtain a PDR fringe frequency distribution diagram; S3, carrying out pixel-by-pixel weighting fusion calculation on the TCP signal characteristic image and the PDR stripe frequency distribution map to obtain a coupling strength distribution map; s4, performing binarization judgment on the ice film area of the ice aggregate detection area according to the statistical parameter of the coupling strength distribution diagram to obtain an ice film mask; s5, generating an aggregate candidate region according to the ultraviolet channel image and the structured light reflection image, and calculating actual size parameters of the ice aggregate in the ice aggregate detection region based on the aggregate candidate region and the ice film mask; s6, carrying out frozen water correction coefficient conversion on the average coupling strength of the coupling strength distribution diagram to obtain the frozen water content of the frozen aggregate detection area; And S7, judging whether the frozen aggregate is qualified or not according to the frozen water content and the actual size parameter. In a preferred embodiment, acquiring an ultraviolet channel image and a structured light reflection image of an ice aggregate detection area comprises: responsive to the detection signal At moment, exciting TCP substances of the ice aggregates in the ice aggregate detection area by an ultraviolet pulse light source to generate luminescence radiation, and collecting an ultraviolet channel image of the ice aggregate detection area; At the position of At the moment of ms, stripe structure light is projected to the ice aggregate through a stripe structure light projector, and a structure light reflection image of an ice aggregate detection area is acquired. In a preferred embodiment, the signal extraction is performed on the ultraviolet channel image to obtain a TCP signal feature image, and the fringe frequency analysis is performed on the structured light reflection image to obtain a PDR fringe frequency distribution map, including: Collecting a background image of the aggregate-free detection area, and performing pixel-by-pixel differential operation on the ultraviolet channel image and the background image to obtain a TCP original differential image; noise cancellation processing is carried out on the TCP original differential image, so that a TCP signal characteristic image is obtained; Performing two-dimensional Fourier transform on the structure light reflection image to obtain a frequency domain amplitude spectrum of the structure light reflection image; searching energy peak coordinate point of frequency domain amplitude spectrum through peak detection algorithm Calculatin