Original Gold identification using Digital Image Processing

Precise determination of metal contents is an important process for consistency check and for getting novel & pure metal objects. Accurate determination become even more important if metal involved are expensive and precious like Gold, Platinum etc.

The method used worldwide for accurate determination of gold and its contents is Cupellation. It is rather a refining process of separating noble metals like Gold from base metal present in the ore. In comparison to this method other techniques are also available which provide better and non-destructive gold contents measurement such as X-Ray Fluorescence (XRF).

XRF testing is relatively simple, fast, reliable, and inexpensive. XRF is based on the principle that atoms are energized by an external energy source and release X-ray photons of a characteristic energy or wavelength. The numbers of photons are counted of each energy from sample. Electronic shell of each atom emits a characteristic radiation under certain conditions. The photons released in the form of energy are compared with the energy of source atoms and in this way the original gold can be detected.

Schematic diagram representing the working of XRF (Image courtesy: research gate)

The un-availability of instruments like XRF spectrometers and Karat meters for XRF testing could be a challenge owing to the cost of set-up involved.

Statistical method based on Digital Image Processing using MATLAB could be a good alternate to this problem for detection of gold. In this method Statistical Parameter Estimation is used where a set of parameters is compared with the predefined parameters and using tolerance matching technique one can identify the original gold objects among several gold objects.

How the Process work?
There are different types of statistical parameters of image like mean, standard deviation, variance, energy etc. There are three colour components of a coloured image red, blue and green. The histograms of each red, green and blue component are plotted and comparison is done on the basis of the intensities of number of pixels present in the image with that of other images. So one image can be differentiated from several other images using correlation between those pixel intensities. Therefore using these parameters one can easily detect the correct image from the given images.

The given block diagram explains the working of this method.

The images of the objects should be stored in the database. Then all the parameters of the image are set in the MATLAB software. It provides the results which are approximately close to the exact results. It is very user friendly and easy to use method.

By – Ms. Manisha Aggarwal – Asst. Prof., ECE, Chitkara University, H.P.

References:

http://iceeot.org/papers/OR0619.pdf, Archaeometry.missouri.edu