Zhang Yujun Guo Yi
(Aerospace Exploration Technology Center, Ministry of Geology and Mining, Institute of Aerospace Exploration)
Abstract: This paper describes the methods and results of displaying, enhancing, and interpreting the aeromagnetic, aerial release, and avionics data of Hami Tudun Survey Area in Xinjiang by using the digital image processing technology. The studied image restoration technique effectively eliminates the strip noise in the aerial release data due to the change of atmospheric background; this method is also applicable to the pre-processing of avionics data. Through the study, three aspects of geological information are extracted by applying three-element image, two-element image, ratio image, scatter map, three-dimensional shadow map, first or second derivative map, local adaptive enhancement map, rose map, grayscale segmentation, K-L transformation, and YIQ-RGB conversion: (1) tectonic traces; (2) lithological filling; and (3) mineral search anomalies. This work shows the charm of digital image processing technology for the method of displaying and interpreting aerial physical exploration data, which has three major features: fast, intuitive and easy to synthesize.
I. Preface
In the field of geosciences, digital image processing was first used mainly for remote sensing. Geophysicists have recognized that any spatially varying geophysical data can be displayed and interpreted using digital image processing techniques. The parameters utilized in this work are: aerial radioactivity (potassium, thorium, uranium, TC total channel), aeromagnetism, tri-frequency avionics (three frequencies, 520Hz, 2020Hz, 8020Hz, real and imaginary components), and guard film false color images as reference.
The highly sensitive integrated aerial measurement system used for the flight consisted of the following equipment: two boxes of NaI crystals with a total volume of 32,000cm3, a proton spinning magnetometer with a sensitivity of 0.5nT and a Tridem electromagnetic system. The flight altitude was 75m and the measurement scale was 1:25,000 meters. Potassium content of the measurement area varies from 0 to 4.6%, thorium content varies from 0 to 47 ppm, and uranium content varies from 0 to 9.1 ppm; the relative dynamic range of the magnetic field is 3540 nT.
The pre-processing of aerial release avionics data
Preprocessing is required for conventional image processing. Due to the instability of the atmospheric background, the raw data of aerial release is often accompanied by the phenomenon of "stripes", and the useful information from the earth is often drowned in the noise of "stripes". In this work, an image restoration technique is investigated, the principle of which is shown in Fig. 1. The principle is shown in Fig. 1, and the method successfully removes the "banding" noise in the aerial data.
Fig. 2 shows the original data image of the total channel in the upper left corner, the noise image obtained by several times of sliding average in the upper right corner, the total channel image after subtracting the noise interference in the lower left corner, and the final recovered image of the total channel in the lower right corner.
The K, Th, and U elemental image is actually a regional geochemical map, which is very similar to the satellite image.
Similar to the aerial release, there are serious "stripes" in the raw avionics data due to the bias value and zero drift of the instrument, and the avionics image has been significantly improved by using the above image restoration technique. Figure 3 (Figure 8 of the color version) shows the comparison between the 8020Hz real component image before and after restoration. Preprocessing of the avionics data also included image editing, Wallis and median filtering.
Figure 1 Processing flow of replay image restoration
Figure 2 Comparison image of replay total channel restoration
Third, the enhancement and interpretation of aerial survey data image
Extraction of tectonic information
Directional conductivity images of the aerial survey and avionics data contain rich geological information, and for aeromagnetism, in order to extract the tectonic information, the most attractive is the stereo shadow image. Stereo shadow image.
Dods et al. (1984) proposed the following formula for calculating the stereo shadow image of magnetic field:
Research Proceedings of Yujun Zhang on New Methods of Geological Surveys
In the formula: λ - the angle between the direction of the light source and the surface normal; φ - the height angle of the light source; θ - the azimuth angle of the light source.
Figure 4 (color version with Figure 8) shows the aeromagnetic color stereo shadow image, which shows both magnetic field magnitude and gradient calculated according to the above equation for both variables, and chooses the light direction to be north-west. The color gradient on each pixel represents the total magnetic field, and the brightness or darkness of the pixel's color varies according to the slope or gradient of the point (Holroyd, 1986).
The two-way derivative map of the magnetic field, the color stereo shadow map, and the image enhanced by local adaptive histogram equalization show the tectonic features of the region. Based on these images, a tectonic trace map was made (Fig. 5, color plate with Fig. 8), and more than 50 tectonic feature lines were identified. The study of tectonic features using the combined interpretation of aeromagnetism, aerial radiography, and aeroelectricity has, in some cases, yielded additional information on the tendency of the section. The rose diagram shows the statistical distribution of the frequency of tectonic shapes.
Lithologic Fill
The RGB-YIQ function is useful for multiparameter image enhancement. It converts the RGB tri-band of a color image to luminance (Y) and chroma (I and Q), approximating YIQ as the equivalent of IHS (luminance, hue, and color saturation). The conversion and inverse conversion are carried out according to the following formulas:
Proceedings of Research Papers on New Methods of Geological Surveys by Yujun Zhang
Proceedings of Research Papers on New Methods of Geological Surveys by Yujun Zhang
In the present survey area, we apply the RGB ←→YIQ transformation to achieve two purposes: ① Utilizing the feature that the correlation among the three components, namely Y, I, and Q, is smaller than that among R, G, and B. The Y, I, and Q components are more correlated with each other. The color purity of the aerial images is improved by RGB-YIQ-SCALE-RGB processing; ② Enhancement of the integrated magneto-electric discharge parameter map. Both maps are useful for lithologic filling.
A six-band image consisting of the amplitudes of three elements of the airborne discharge (K, Th, and U) and three frequencies of the airborne electricity was obtained by unsupervised classification of nine lithological classes (Fig. 6): (i) ultramafic rocks; (ii) Cu-Ni ore targets; (iii) sand targets; (iv) granites; (v) syenites; (vi) metamorphic rocks; (vi) mixed rocks; (viii) quaternary sediments; (ix) Tertiary and Quaternary deposits; and (x) Tertiary deposits.
The mean vectors (K, Th, U) of these categories are listed in Table 1 below.
Table 1
Extraction of Cu-Ni and alluvial anomalies
Based on the pattern of known Cu-Ni ores with low K, located near the fracture and accompanied by localized magnetic anomalies, 20 Cu-Ni anomalous targets were circled, and two of them were confirmed as Cu-Ni-rich anomalies by ground Two of these anomalies have been confirmed to be Cu-Ni rich mineralization by ground work, with one grading up to 2% Cu. The average K content of these anomalies is 0.84%.
The alluvial anomalies are defined by high Th and U values and are classified as quaternary sediments with average values of 15.4 ppm Th and 6.09 ppm U. The high anomalies in the Th and U channels are likely to be caused by zircon and monazite.
Originally published in Exploration Geophysics Beijing (89) International Symposium (UCEGSEG), Abstracts of Papers, 1989.