CN-121982553-A - Intelligent detection method and system for activity parameters of desulfurization strains based on hyperspectral images

CN121982553ACN 121982553 ACN121982553 ACN 121982553ACN-121982553-A

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a method and a system for intelligently detecting activity parameters of desulfurization bacteria based on hyperspectral images, wherein the method comprises the steps of extracting original reflectivity of each pixel point in hyperspectral images of a desulfurization bacteria liquid under a continuous wave band; the method comprises the steps of anchoring a local background baseline by utilizing original reflectivities at a first reference wave band and a second reference wave band, determining metabolic absorption depth by combining the original reflectivities at a characteristic metabolic wave band, determining a spatial spectrum gradient bias quantity based on the original reflectivities of each pixel point and the spatial neighborhood pixel points of the pixel points at a structural scattering wave band and the metabolic absorption depth difference between the pixel points, and determining strain activity evaluation parameters according to the spatial spectrum gradient bias quantity and the metabolic absorption depth of all the pixel points. The invention can eliminate the interference of background water fluctuation and high-density inactivated bacteria liquid, realize the real characterization of the microorganism metabolism intensity and improve the monitoring precision and stability of the desulfurization system.

Inventors

HU MINGLIANG
RUAN YANGYANG
ZHANG YANG
CHEN CONG
LI SHULONG

Assignees

武汉谦豫科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. The intelligent detection method for the activity parameters of the desulfurization strains based on hyperspectral images is characterized by comprising the following steps of: Extracting the original reflectivity of each pixel point in the hyperspectral image of the desulphurized bacteria liquid under the continuous wave band; Determining a characteristic metabolism wave band, a first reference wave band and a second reference wave band from the continuous wave band, anchoring a local background baseline by utilizing the original reflectivities of each pixel point at the first reference wave band and the second reference wave band, and carrying out ratio bias operation by combining the original reflectivities at the characteristic metabolism wave band to obtain a metabolism absorption depth; Determining a structural scattering wave band from the continuous wave band, constructing attenuation weights based on the original reflectivity difference of each pixel point and the spatial neighborhood pixel point at the structural scattering wave band, and carrying out weighted summation on the metabolic absorption depth difference among the pixel points by using the attenuation weights to obtain a spatial spectrum gradient bias; And constructing a variation coefficient modulation item according to the mean value and standard deviation of the spatial spectrum gradient bias values of all the pixel points, weighting the metabolic absorption depth by using the spatial spectrum gradient bias values of all the pixel points, and fusing the weighting result with the variation coefficient modulation item to obtain a strain activity evaluation parameter.
2. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization bacteria according to claim 1, wherein determining a characteristic metabolic band and first and second reference bands comprises: The method comprises the steps of comparing a high-activity desulfurizing bacterium sample with an inactivated desulfurizing bacterium sample, calibrating a section where an original reflectivity minimum value caused by cytochrome absorption photons in a spectrum curve of the high-activity desulfurizing bacterium sample is located as an empirical range of a characteristic metabolism wave band, searching a wave band section which is not influenced by the activity state of the desulfurizing bacterium obviously and has the most gentle change of the original reflectivity at the left side and the right side of the characteristic metabolism wave band, calibrating the wave band section as an empirical range of a first reference wave band and a second reference wave band respectively, calculating average original reflectivity of each wavelength in a hyperspectral image of the desulfurizing bacterium liquid, determining the wavelength with the smallest average original reflectivity in the empirical range of the characteristic metabolism wave band as the characteristic metabolism wave band, and correspondingly determining the wavelength with the largest average original reflectivity as the first reference wave band and the second reference wave band in the empirical range of the first reference wave band and the second reference wave band respectively.
3. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization bacteria according to claim 1 or 2, wherein the metabolic absorption depth satisfies the expression: ; In the formula, Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 Metabolic absorption depth of pixel points of the columns; Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 The pixel points of the columns are in the characteristic metabolism wave band Original reflectivity at the point; Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 The pixel points of the columns are in a first reference wave band Original reflectivity at the point; Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 The pixel points of the columns are in the second reference wave band Original reflectivity at the point; Indicating a first prevention zero removal minimum constant.
4. The method for intelligently detecting activity parameters of a desulfurization bacterial strain based on hyperspectral images according to claim 1, wherein the step of determining a structural scattering band from continuous bands comprises the steps of: Extracting spectral data of desulfurization bacteria suspensions with different physical densities in a near infrared band, calibrating a continuous spectral interval in which the correlation coefficient between the original reflectivity and the physical densities reaches a forward maximum confidence interval and the original reflectivity difference and the adjacent wavelength variation gradient between different active samples under the same physical densities are minimum as an empirical range of a structural scattering band, calculating the average original reflectivity of each wavelength in a hyperspectral image of the desulfurization bacteria liquid, and determining the wavelength with the maximum average original reflectivity in the empirical range of the structural scattering band as the structural scattering band.
5. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization bacteria according to claim 1 or 4, wherein the spatial spectral gradient bias satisfies the expression: ; In the formula, Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 Spatial spectral gradient bias of the pixel points of the columns; The first image is expressed in hyperspectral image of desulfurization bacteria liquid Line 1 A spatial neighborhood with the pixel points of the columns as the center; representing spatial neighborhood Row and column coordinates of any one neighborhood pixel point in the image; 、 Respectively represent the first of hyperspectral images of the desulphurized bacteria liquid Line 1 Pixel points of columns, neighborhood pixel points Is a metabolic absorption depth of (2); 、 Respectively represent the first of hyperspectral images of the desulphurized bacteria liquid Line 1 Pixel points of columns, neighborhood pixel points In the structural scattering band Original reflectivity at the point; An exponential function that is based on a natural constant; indicating a second prevention zero removal minimum constant.
6. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization strains according to claim 1, wherein the activity evaluation parameters of the strains satisfy the expression: ; In the formula, Representing strain activity evaluation parameters; And Respectively representing the height and the width of the hyperspectral image of the desulphurized bacteria liquid; Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 Metabolic absorption depth of pixel points of the columns; Represents the first step in hyperspectral image of the desulphurized bacteria liquid Line 1 Spatial spectral gradient bias of the pixel points of the columns; The standard deviation of the spatial spectrum gradient offset of all pixel points in the hyperspectral image of the desulphurized bacteria liquid is represented; the average value of the spatial spectrum gradient offset of all pixel points in the hyperspectral image of the desulphurized bacteria liquid is represented; Representing natural constants; A logarithmic function that is based on a natural constant; indicating a third prevention zero-divide minimum constant.
7. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization bacteria according to claim 1, further comprising: generating a comprehensive control threshold based on the historical strain activity evaluation parameters in the time sliding window, comparing the strain activity evaluation parameters at the current moment with the comprehensive control threshold, and outputting a corresponding process control instruction to the automatic control system according to the comparison result.
8. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization strains according to claim 7, wherein the generation of the comprehensive control threshold based on the historical strain activity evaluation parameters in the time sliding window comprises the following steps: The method comprises the steps of obtaining the current hydraulic retention time of a sulfur-containing wastewater bioreactor, establishing a time sliding window by taking the current hydraulic retention time as a time span, extracting all strain activity evaluation parameters which are continuously output in a history in the time sliding window, obtaining the mean value and standard deviation of all strain activity evaluation parameters in the time sliding window, calculating the difference between the mean value of the strain activity evaluation parameters and the standard deviation of the strain activity evaluation parameters which are 3 times, taking the difference as an adaptive activity dynamic threshold, and extracting the maximum value of the adaptive activity dynamic threshold and a preset absolute activity baseline threshold as a comprehensive control threshold.
9. The hyperspectral image-based intelligent detection method for activity parameters of desulfurization strains according to claim 7, wherein the outputting of corresponding process control instructions to an automatic control system according to the comparison result comprises: And responding to the fact that the activity evaluation parameter of the strain at the current moment is smaller than the comprehensive control threshold, judging that the desulfurization bacteria in the current sulfur-containing wastewater bioreactor are deactivated abnormally, generating a low-activity early warning signal, and outputting a process control instruction for increasing the dosage of the nutrient and the opening frequency of the bottom sludge discharge valve to the automatic control system.
10. The hyperspectral image-based intelligent detection system for the activity parameters of the desulfurization bacteria species is characterized by comprising a processor and a memory, wherein the memory stores computer program instructions which, when executed by the processor, realize the hyperspectral image-based intelligent detection method for the activity parameters of the desulfurization bacteria species according to any one of claims 1 to 9.

Description

Intelligent detection method and system for activity parameters of desulfurization strains based on hyperspectral images Technical Field The invention relates to the technical field of image processing. More particularly, the invention relates to a hyperspectral image-based intelligent detection method and system for activity parameters of desulfurization strains. Background In the industrial sulfur-containing wastewater treatment process, the life activity of the desulfurization strain is a core element for determining the treatment efficiency of a biological desulfurization system and the stability of the effluent quality, so that the life activity is important for monitoring the state of flora in a reactor in real time. Currently, hyperspectral imaging technology has been introduced into monitoring of desulfurization bacteria by virtue of advantages of non-contact and high resolution, and the prior art generally uses a hyperspectral camera to collect images of bacteria liquid in a bioreactor, and utilizes an original reflectivity average value under a specific characteristic wave band, or extracts reflectivity of a characteristic absorption wave band and a near infrared reference wave band and calculates static ratio of the reflectivity to the reflectivity of the characteristic absorption wave band and the reflectivity of the near infrared reference wave band, so as to evaluate the overall concentration and the activity state of mixed bacteria liquid in the desulfurization reactor. However, the environment of the industrial bioreactor which is actually operated is complex, a large amount of insoluble elemental sulfur particles can be accumulated on the outer surface of cells of high-activity desulfurization bacteria in the process of metabolizing and decomposing sulfides in wastewater, the accumulation of the biological metabolites can obviously change the optical scattering characteristics of single bacterial colonies and surrounding microenvironments, when a large amount of high-density inactivated desulfurization bacterial colonies exist in a wastewater treatment system, the macroscopic optical scattering intensity of the whole high-activity desulfurization bacteria is very similar to that of a small amount of high-quality desulfurization bacterial colonies which are in high activity, the prior art only depends on a single reflectivity average value or a fixed global spectral index of a full-view area, the characteristic of spatial spectral heterogeneity of the high-activity desulfurization bacterial colonies on microscopic spatial distribution, which is generated by uneven aggregation of metabolites, is ignored, so that a monitoring system frequently misjudges the high-density inactivated bacterial liquid into a high-activity metabolic state, thereby outputting distorted feedback signals to an automatic control system, not only can not provide accurate nutrition dosing guidance for the process, but also can easily trigger the collapse of the desulfurization system, the wastewater treatment quality is seriously affected, and the effluent is difficult to reach standards. Disclosure of Invention In order to solve the technical problem that the high-density inactivated bacterial liquid is misjudged to be in a high-activity state due to the fact that the spatial spectrum heterogeneity generated by aggregation of high-activity flora metabolites is ignored in the prior art, the invention provides the scheme in the following aspects. In a first aspect, the invention provides a hyperspectral image-based intelligent detection method for activity parameters of desulfurization strains, which comprises the following steps: extracting original reflectivities of all pixel points in a hyperspectral image of a desulphurized bacteria liquid under a continuous wave band, determining a characteristic metabolism wave band, a first reference wave band and a second reference wave band from the continuous wave band, anchoring a local background baseline by using the original reflectivities of all pixel points at the first reference wave band and the second reference wave band, carrying out ratio offset operation by combining the original reflectivities at the characteristic metabolism wave band to obtain metabolism absorption depth, determining a structure scattering wave band from the continuous wave band, constructing attenuation weights based on the original reflectivities of all pixel points and the spatial neighborhood pixel points, carrying out weighted summation on the metabolism absorption depth differences among the pixel points by using the attenuation weights to obtain a spatial spectrum gradient offset, constructing a variation coefficient modulation item according to the average value and standard deviation of the spatial spectrum gradient offset of all pixel points, weighting the metabolism absorption depth by using the spatial spectrum gradient offset of all pixel points, and fusing a weighting result with th