| Abstrakt: |
The R-value image mapping method related to ore atomic number can identify ores more accuratelywhen applied to mineral identification, but owing to the deduction of R-value directly ignores the influence of orethickness, it is greatly affected by the thickness, so this image mapping method has certain defects. The T-value algorithm related to the thickness and density of ore retains the two important parameters, therefore, using T-value will obtained more image information. T-value image mapping method is used to verify the acquireddual energy X-ray transmission image of ore. First of all, advantages and disadvantages of the curve fittingbetween T-value and R-value recognition algorithm are compared. Through experiments, it can be seen that thecorelation coefficients of R-value curve fitting were 0.900 24 and 0.880 31, while the correlation coefficients of T-value curve fitting reach 0.998 4 and 0.991 16 respectively. Therefore, the T-value recognition algorithmrelatively reduces the influence of thickness on material recognition and the T-value algorithm is better than R-value algorithm. Therefore, an image mapping method based on T-value is proposed on the basis of this research. In which the intensity of grayscale substitution is adopted to calculate the T-value, which is then used asthe equal size image as the grayscale, and then the mathematical morphology image processing and median filtertechnology are used successively to process the image sharpness, and finally RGB pseudo-color processing is usedto obtain the final image. According to the experimental result, The T-valued image mapping method not onlyimproves the running speed compared with the R-valued image mapping method, but also better in enhancing theimage clarity and ore recognition accuracy. [ABSTRACT FROM AUTHOR] |
