THE CONTRIBUTION OF REMOTE SENSING AND AEROMAGNETISM TO GOLD PROSPECTING: THE CASE OF THE MEIGANGA ZONE, CAMEROON

In order to optimize gold prospecting in the Meiganga zone located in the Adamaoua region of Cameroon, aeromagnetic and remote sensing prospecting was carried out in the eastern and southern parts. The remote sensing approach on a Landsat 8 OLI/TIRS image highlighted areas of maximum gold concentration. positive anomalies. The grid of the reduced residual equatorial anomaly (ARRE) confirms that the local geology is strongly magnetic (gneiss and quartzite). The filters of the derivatives allowed to establish a map of magnetic lineaments of major orientation N045° and minor orientation N130°. The horizontal gadient superimposed on the local maxima showed the presence of deep structures oriented NE-SW. The analytical signal superimposed on the local maxima highlights the metamorphic basement consisting of rocks with strong magnetism. The application of Euler deconvolution localizes the depth of the sources of linear anomalies.

In order to optimize gold prospecting in the Meiganga zone located in the Adamaoua region of Cameroon, aeromagnetic and remote sensing prospecting was carried out in the eastern and southern parts. The remote sensing approach on a Landsat 8 OLI/TIRS image highlighted areas of maximum gold concentration. Thus, ferric ion bearing minerals are located in the North-West, silicate minerals bearing ferrous ions are in the Centre while clay minerals are in the North-East and East. The principal component analysis revealed important structural information. The PCA Spatial Map (PC1, PC2, PC3) showed the plutonic formations composed of anatexis and anatexis granites, vegetation cover (at the date of image acquisition: February 22, 2019), areas of permanent water circulation or accumulation, and metamorphic and sedimentary formations namely gneisses, quartzites, schists and superficial clay formations. A Landsat SRTM (Shuttle Radar Topography Mission) image was also used to enhance the lineaments through the Sobel filter to highlight the geomorphological (cliffs, valleys, ...) and topographic (river network, ridge and drainage segment) structures. The aeromagnetic approach was also important. The study of the modified magnetic field (CM) showed 4 ranges: very high, high, medium and low. The Total Magnetic Anomalies (TMI) of the area are subdivided into 2 ranges; large positive anomalies (221.1-103.0 nT) located in the lower part of NE-SW orientation, small positive anomalies (103.0-(-)89.7 nT) located in the upper part of NE-SW orientation. The reduced total magnetic anomaly at the equator shows a fairly similar distribution to the total magnetic anomaly with the large positive anomalies in almost the entire lower part. Superimposed on the geological map, "Neoproterozoic pre-to syntectonic granitoids (C)" are superimposed on the large positive anomalies and "Neoproterozoic conglomerates, quartzites, sedimentary shales and volcanosedimentary rocks (A)" and "Neoproterozoic syntectonic granitoids (B)" are superimposed on the large and small 776 positive anomalies. The grid of the reduced residual equatorial anomaly (ARRE) confirms that the local geology is strongly magnetic (gneiss and quartzite). The filters of the derivatives allowed to establish a map of magnetic lineaments of major orientation N045° and minor orientation N130°. The horizontal gadient superimposed on the local maxima showed the presence of deep structures oriented NE-SW. The analytical signal superimposed on the local maxima highlights the metamorphic basement consisting of rocks with strong magnetism. The application of Euler deconvolution localizes the depth of the sources of linear anomalies.
Gold is a precious resource and is exploited artisanally or semi-mechanically in Cameroon, but only to a limited extent (Asaah, 2010; Manga et al., 2017). By combining remote sensing and aeromagnetic methods (Pour et al., 2013; Pour and Hashim, 2015a), it is essential to detect zones favourable to gold mineralization in the locality of Meiganga. Since gold is diamagnetic, the presence of minerals such as oxides, sulphides and clay minerals can be used as indicators for the location of hydrothermal alteration zones associated with gold occurrences (Poulsen, 1996). .The objective of this work will be to process and interpret satellite images (Landsat 8 OLI / TIRS and Landsat SRTM) and aeromagnetic maps in order to correlate the observed magnetic anomalies, geology, structural and gold (placer) work site data in the study area and to produce a favorability map at the end.

Presentation of the study area General context
Located in Cameroon in the Adamaoua region, Mbéré department and Meiganga district, our study area is the area delimited by three aeromagnetic sheets crossed by the N°1 with an area of about 1575 km2.
Geomorphologically speaking, our study area belongs to the Adamaoua plateau, located in the middle of the savannah, where the climate is tropical with two seasons, a dry and a rainy one. This locality is characterized by a relief with a convex-concave morphology separated by U and V-shaped valleys or Talwegs (Suchel, 1987). The hydrographic network of the study area is dense and dendritic, with the main collector being the LomRiver ( Figure  1A).
Our study area is the area bounded by three aeromagnetic leaves (NB-33-XV-11, NB-33-XV-14, NB-33-XV-15) from CAPAM. This zone is located south and east of the Meiganga district and borders the Central African Republic on its eastern part ( Figure 1B).

Geological setting
The zone belongs to the granitic and polyclinic Adamawa basement which is a polycyclic complex with a general N130° orientation that extends to the Central African Republic (Ganwa, 2005;Tchameni et al. (2006). This shows that this old complex contains mainly alkaline biotite granites, anatexis granites, biotite gneisses, shales, and quartz and pegmatitic veins (Guiraudie, 1955;Humbell, 1966).
Meiganga is located in the central part of the Pan-African fold belt (Ganwa, 2005), which is a megatectonic structure formed during the Neoproterozoic and is the product of collision between the Saharan metacraton and the   Table 1 shows the characteristics of the Landsat-8 satellite. Landsat-8 data have a high signal-to-noise ratio and 12-bit quantization of the data that permit measurements of subtle variability in surface reflectance (Irons et   Data pre-processing For this study, different softwares were used: Erdas Imagine and ENVI: which are complete commercial remote sensing software applications capable of preprocessing, enhancing, transforming, and classifyingremote sensing images to extractspatial and spectral information related to geology, such as lithology, hydrothermal weathering, structure, etc.; ArcGis : designed by ESRI, is a suite of software (ArcMap, Arcglobe, ArcCatalog and ArcScene) that has been used for digitizing and joining Excel files; CPI Geomatica: has excellent modules for extracting structural elements; Rose.Net: allowed the graphical representation and calculation of the statistical distribution of the orientation of lineaments (faults, dykes, etc.); Rock Works allowed the synthesis of geological information and geomodelling in rosette.
The pre-processing consisted in correcting the geometric and radiometric distortions (errors) of the platforms and sensors. As for our original data, they underwent three (03) corrections, namely atmospheric correction, noise reduction and data resizing (Cooley et al., 2002). Atmospheric correction was performed through radiometric calibration, mathematical correction and dark object subtraction; Concerning noise reduction, we proceeded with MNF (Minimum / Maximum Noise Fraction) transformations (Research Systems, Inc, 2008). These results are derived from the processing of remote sensing datasets from our study area, and are presented in this section under two aspects: the index aspect, which is intended to specify zones of high gold concentrations (based on hydrothermal alteration representing the surface expression of gold-bearing sulphide and silicate deposits); and the lithological and structural aspect, which could be used to detect related mineral deposits.

Coloured compositions
After all the pre-processing operations performed on the original spectral bands, the "true colour" display ( Figure 3) is obtained by performing an RGB colour composite (432); in fact, the image obtained is the one that the eyes would 780 be able to observe if they were in the place of the satellite sensor. However, the spectral bands of this image are highly correlated, so the resulting image is not very easy to interpret visually. To this end, the results of the IFO calculation allow us to establish the colour compositions between the least correlated bands. The results of the IFM calculation allowed us to obtain the following   The interpretation of the three (03) spaciocards (mentioned above) depends on the grey level: the lightest areas are the areas of highest concentration (in minerals corresponding to the band ratio). Following the example of the spaciocard resulting from the 4/2 ratio (appendix 1 figure 19), the minerals (ferric ion carriers (hematite, magnetite, ilmenite, rutile,...) are more concentrated in the Northern, North-Eastern and South-Eastern parts of our study area (Inzana et al., 2003).

781
By combining the three RGB band ratios (R:4/2, V:5/6, B:6/7), the spaciocard presented in Figure 5 is obtained, which highlights the areas of maximum concentration for the different band ratios used. For this purpose, we have the following interpretation table 3.  Figure 5, the presence of the violet colour illustrates the zones with the highest gold concentration, which are indeed the zones that have been most affected by the hydrothermal alteration. Violet is the colour resulting from the mixing of the red and blue colours(see corresponding minerals in table 4).

Principal Component Analysis (PCA)
To this end, PC1 (Appendix 1, Figure 21) highlights features common to all input bands and often displays important structural information. PC2 (Appendix 1, Figure 21) is orthogonal to PC1 in directional space and highlights visible spectral differences and spectral bands. PC3 (Appendix 1, Figure 21

Spatial filtering
The mapping of geological lineaments is important for mineral exploration because of the high potential of these lineaments to shelter orebodies; orebodies that are transported anddeposited byascending hydrothermal fluids (Pour  et al., 2018c). For the geological and geomorphological linear delineation, we used the first main component (PC1) of the PCA (Principal Component Analysis) previously performed, as this component contains most of the nonredundant information (more than 90%). The enhancement of the lineaments which was carried out through the Sobel filter (Appendix 1 figure 22): after analysis, the lineament map resulting from this filter (Sobel) mainly highlights the geomorphological (cliffs, valleys, ...) and topographic (hydrographic network, ridges and drainage segments).

782
Aeromagnetic data processing Pre-processing For this work, we were provided with three aeromagnetic maps in ".PDF" format. These 1/50 0000 scale maps were produced during the aeromagnetic survey of Cameroon carried out by a Canadian Company. This Gold-Cameroon project was carried out in 1986 in collaboration between the Cameroonian government and BRGM. These aeromagnetic data were processed using Geosoft Oasis Monaj 8.4 software; topographic, geological and structural data were processed using ArcGis 10.5 software.
We scanned aeromagnetic maps from ".pdf" to ".jpeg" format using Adobe Acrobat software and then georeferenced the ".jpeg" files on ArcGis. The continuation of our work consisted in the digitization of isomagnetic curves of the georeferenced images to obtain at the end a map of the isomagnetic curves (Appendix 1 figure 17).
Aeromagnetic data pre-processing operations were carried out using ArcGis software. The lines obtained after digitizing are converted into points using the 'Generate point along line' tool with a deviation of 0.0001°, then the attribute table of the layer thus created is accessed and the longitude/latitude of the points is calculated in decimal degrees. At the end of this operation, we obtained a ".shp" file whose attribute table gave us information on the intensity of the modified magnetic field, the longitude and latitude of each point generated.

Modified magnetic field (CM) and theoretical geomagnetic field
The map of the modified magnetic field (Appendix 1 Figure 23 The reference geomagnetic field map of the locality at the time of the magnetic measurements (Appendix 1 Figure  24) is a function of magnetic tilt and declination. This filter is calculated by choosing inclination and declination values from the International Geomagnetic Reference Field (IGRF) normal field model at a time period corresponding to the period of data acquisition.

Total Magnetic Anomalies (TMI)
This map (Figure 7) represents the distribution of magnetic anomalies (Miller and Singh, 1994) in the Study Area. We can therefore subdivide these values into two ranges of anomalies: large positive anomalies (221.1-103.0 nT): represented by the colours orange, magenta and red. These anomalies are located in the lower part of the map and represent areas where the magnetic intensity is higher than the theoretical magnetic field value. They are oriented North-East and East; small positive anomalies (103.0 -(-)89.7nT): represented by the colours yellow, green and blue. They are located in the upper part of the map and represent the areas where the magnetic intensity is higher than the theoretical magnetic field value. They are oriented East and North-East.

Equator reduction and residual anomaly
The map of the reduced total magnetic anomaly at the equator (Appendix 1, Figure 25) has a fairly similar distribution to the previous map with the difference that the large positive anomalies occupy almost the entire lower part of the map.
The reduced equatorial anomaly map was superimposed with the geological map of the Study Area (Appendix 1 Figure 26) to better understand the distribution of the anomalies. Note that the "Pre-and syn-tectonicNeoproterozoicgranitoids (C)" are superimposed on the large positive anomalies while the "Conglomerates, quartzites, sedimentary and volcanic schists (A)" and the "Syn-tectonic Neoproterozoicgranitoids (B)" are superimposed on the large and small positive anomalies.
The residual anomaly grid (ARRE) is obtained by subtracting grids (TMI -AR), (El Gout et al., 2009). The "reduce to magnetic equator" filter is applied to the result to obtain the grid (Figure 8) of the reduced residual anomaly at the equator (ARRE).
Local anomalies are ranged from -128.8 nT to 168.3 nT.Comparing this map with the reduced total anomaly map at the equator (Appendix 1, Figure 25), it can be seen that the magnetic units corresponding to the positive anomalies 783 (magenta, red, orange, yellow colours) have increased significantly in the lower part of the map but have decreased in the upper part of the map. These variations confirm that, in general, the local geology of our study area consists essentially of highly magnetic rocks such as gneisses and quartzites (Debeglia, 2005).

Figures 7 and 8:-
Map of the total magnetic anomaly, Map of reduced residual anomalies at the equator (ARRE).

Derivative filters
We performed horizontal derivatives (following X and Y) and a vertical derivative (following Z) on the grid (TMI_R), (Salem et al., 2008). However, the aim is to highlight the magnetic lineaments perpendicular to the X plane. Thus the analysis of the map (Figure 9) of the horizontal derivative (following X) allowed us to establish a map of magnetic lineaments of the locality (Figure 10).
This method, used by several authors (Everaerts and Mansy, 2001; Abderbi and Khattach, 2011), allows on one hand to locate the zones with rapid variation of the magnetic field caused by the lithological change or the presence of geological discontinuities (fractures, faults…), and determine the dip direction of thoses geological structures on the other hand. Indeed, the magnetic anomalies correspond to inflection points which transform after the horizontal gradient calculation into local maxima.These maxima are located above the geological contacts that present magnetic susceptibility contrasts (Van Senden et al., 1990). In order to determine the dip direction of thoses structures, a series of upward continuation has been carried out at different altitudes. For each level the horizontal gradient of the residual magnetic field is calculated and its local maxima are determined. If the structures are vertical, the maxima obtained at each altitude are superimposed.

Horizontal gradient (GH) and analytical signal
On the map of the horizontal gradient superimposed on the local maxima (Appendix 1, Figure 27), it can be seen that the straight contacts are more pronounced in the central and upper parts. There is also a preferential alignment of the maxima in the SW-NE direction. These results suggest the presence of deep structures oriented in this direction.
The map of the analytical signal superimposed on the local maxima (Appendix 1 Figure 28) shows only positive anomalies, unlike the map of the horizontal gradient (Appendix 1, Figure 27), which shows both positive and negative anomalies. This variation could be due to the presence of a metamorphic basement consisting of rocks with strong magnetism (gneiss, quartzite).
784 Figures 9 and 10:-Horizontal derivative map according to X; surface magnetic lineaments obtained by interpretation of the horizontal derivative map according to X.

Euler's deconvolution
The application of Euler's deconvolution requires knowledge of four parameters (Reid et al., 1990): the structural index, the size of the filtering window, the tolerance and the depth of investigation. After several tests, the best model of Euler's solutions was obtained by fixing a structural index of 1.5, a filtering window of 10, a tolerance of 20%. The Euler solution grid illustrates the position and depth of the sources of linear anomalies. In order to better understand this distribution, we superimposed on this grid the maxima of the horizontal gradient and the maxima of the analytical signal (Figure 11).
By the Blakely and Simpson method, the interpretation of the result of this superposition is based on three criteria: when the maxima of the horizontal gradient are isolated on the map of the two superimposed methods, they represent the true contacts; when the maxima of the analytical signal and of the horizontal gradient are quasi-parallel and not confused, then the analytical signal represents the true contacts and the horizontal gradient indicates the direction of dip of these contacts; and when the maxima of the two methods are confused, then they represent the vertical contacts.The result of this interpretation is a synthetic map that illustrates the distribution of deep faults in the locality (Figure 12). The geological (direction and direction of dip) and geometric (approximate depth and length) characteristics of the deep faults are presented in table 4 in Appendix 2; Table 5 shows activities of CAPAM. 785

Results and Discussion:-
Based on the interpretations of the satellite images, the principal component analysis allowed us to identify several types of geological formations in our study area based on colour. By comparing this PCA map with the geological map of the area, a correlation of the geological formations can be seen (Figure 13). By superimposing the lineaments obtained by Sobel 7.7 filter (at high density) on our PCA map (Appendix 1 figure  29), we can see that these lineaments completely cover the yellow formations. This shows once again that these 786 formations are located at the level of major accident zones such as faults, fractures, river circulation zones, hence the alteration zones of hydrothermal deposits favourable for gold mineralization (Ganwa, 2005.,Ganwa et al., 2016).
By superimposing on the geological map the data of the gold workings, the major superficial lineaments of the Sobel 7.7 filter, the deep Euler faults and the superficial magnetic lineaments of the horizontal derivative X we obtained the following synthesis map in figure 14.
The data from alluvial and eluvial gold workings are much more concentrated on formations of metamorphic and sedimentary types, i.e., quartzites, conglomerates, sedimentary schists and volcano-sedimentary shales. We can also notice by correlation of the structures (deep faults and superficial lineaments) highlighted on our map that the superficial magnetic lineaments of the horizontal derivative X fit exactly several times with the tectonic lines of the existing geological map. We can deduce that the structures obtained have a good accuracy. However, in order to confirm and make these results reliable, it will be necessary to make a field descent and do the detailed mapping. In addition, the favourability map (Figure 15) based on the structural synthesis map already gives us a fairly precise knowledge of the zones favourable for primary gold mineralization in our study area.

Conclusion:-
The identification of potential gold sites using a combination of remote sensing methods and aeromagnetism to detect zones of maximum gold concentration in the Meiganga locality is promising. We have shown the correlation between the existing geological formations and those obtained from remote sensing methods, with emphasis on the alteration formations of hydrothermal deposits, and we have created a favorability map containing all the structures obtained from the processing of aeromagnetic data and satellite images.
Based on these results, we can already say, subject to detailed mapping, that the lineament and fault overlay zones as well as the zones of metamorphic-type formations are likely to contain a primary source of mineralization.