| Scale number is a non-linear definition of light transmission of light-controlled goggle filters,and the scale number tester is a testing device for light transmission.With the increase in the use of light-controlled goggles,the detection of their performance indicators has also gained attention.The existing research on the performance index detection of light-controlled goggles is mostly focused on the response time and retention time of light-controlled goggles,while the research on scale number is not perfect.The existing methods of scale number detection are mainly formula method and spectrophotometer measurement,and only the original data are homogenized.This article proposes a design method for optical path optimization,front and rear sensor driver circuits,signal conditioning circuits,and humanmachine interfaces;Based on the data collected by the optical goggles color number detector system,a method of using data mining technology to detect color numbers was proposed,and modeling analysis was conducted.In the light-controlled goggle scale number detector system,a zero-bias operation mode sensor driver circuit is used,and the active filtering in the signal conditioning circuit and the ADS114S08 high-precision A/D converter with trap filtering are combined to realize the filtering processing and digital acquisition of weak signals.The programmable gain characteristic of this A/D converter is used for program-controlled amplification to solve the problem of large dynamic range transmittance detection of different scale number.In the process of data processing,the original transmittance data was corrected to solve the influence of the environment and the components’ own offset on the signal acquisition;the corrected data was used to compare and analyze with the theoretical transmittance of the standard filter to verify the good linearity and precision of the color number detector system.Through further analysis of the data,a feature description vector is constructed in order to mine more information of the data.The training set and test set based on the feature description vector are constructed by processing the test data of a large number of standard filters.Based on the support vector machine(SVM)model,the gray wolf optimization algorithm(GWO),whale optimization algorithm(WOA)and particle swarm optimization algorithm(PSO)are used to optimize their parameters through the training set,respectively,among which the detection accuracy of the SVM model optimized by the particle swarm optimization algorithm is the highest.By optimizing the design of the light-controlled goggle scale number detector system to complete the accurate collection of light transmission data,and the use of PSO-SVM model for scale number prediction,the results show that: the light-controlled goggle scale number detector based on data mining technology,can meet the needs of scale number detection. |
