Multispectral Imaging and Monitoring of Gold Mine's Using Quadcopter Camera - Coursework Example

Title page Table of Contents Title page 1 Table of Contents 2 Introduction 3 Multispectral imaging 4 Sampling 4 Multispectral imaging using infrared 5 Data representation 6 Multispectral imaging using visible light 7 How to get the ultraviolet illumination 8 Reflected images 9 Image analysis 10 How to produce and receive UV radiation images 10 UV reflected images 11 Radiation and source filters 12 Conclusion 13 References 15 Introduction The increasing demand for gold in the market has resulted in development of economic geology which focuses on recording the mode of occurrence of their deposits. Remote sensing using hand held cameras were used for the first time in 1940s to explore minerals. This technique later progressed to produce gray shade through color aerial photo to a more complex space technique multispectral digital imaging systems and satellite. Multispectral remote sensing has been successful in producing detailed data about mineralogy of various types of rocks in the surface of the earth (Gabr, Ghulam & Kusky, 2010). Remote sensing involves acquiring information about a scene or an object. The color and grayscale imaging systems has dominated this field in visible range of electromagnetic spectrum. Longwave infared imaging that is similar to grayscale imaging does not rely on light reflection to produce an image, but it utilizes thermal emission of an object (Slonecker et al., 2010). Multispectral imaging is the process used to make observation of objects using a range of wavelengths in electromagnetic spectrum which goes beyond the capabilities of the eye (Dyer, Verri & Cupitt, 2013). Recent development in imaging has developed to include not just three color bands, but to include many bands that covers visible spectrum, short wave infrared and near infrared bands, which are represented by SWIR and NIR bands respectively. This development is made possible by advances in focal plane technology that exploits the fact that the materials consist of objects that absorb, reflect, emit and scatter electromagnetic radiation which is dictated by their shape, and their molecular constituents. The radiation that arrives at the sensor is measured at different wavelengths over a wide spectral band, and the spectrum produced is used to identify the materials in the scene (Moon, Whateley & Evans, 2009; Slonecker et al., 2010). This report focus on the mapping high potential zones for gold which can be obtained using modified camera used in the quadcopter, which relies on silicon based sensors that covers between 350nm and 1700nm. The cameras that can record infrared radiation with radiation of between 700 and 1700 nm are based on InGaAs sensors, but this technology is often expensive. For this reason, the performance of the existing technologies used in image spectra and gold extraction in n-dimension spectral space is explored (Dyer, Verri & Cupitt, 2013). Although the measurements wavelength reveals more information about the object, the image obtained may not be interpreted easily. It requires imagery processing to remove all the important information in the spectral bands. Multispectral imaging Every object on the surface of the earth can reflect light in unique patterns; depending on the manner in which light of various wavelengths are absorbed or reflected from the each object. The reflected light is filtered to form unique wavelengths of the electromagnetic spectrum images for different materials. Multispectral imaging utilizes wavelength band in and outside the visible band of the spectrum. Sampling The collection of spectral image information involves operations that include radiometric, spectral and spatial data imaging. Spectral sampling can be realized through decomposition of radiance received from the spatial pixels into different wavebands that varies depending on their resolution. These bands may overlap depending on the type of sensor. A colored image consists of blue, red and green bands in which the spectral bands do not overlap. Digital data is obtained from the conversion from analog to digital. The data obtained is three dimensional spectral cubes. The figure below shows that way in which scan lines can be stacked to produce a 3-dimension spectral data cube with information in x, y and z dimensions (Slonecker et al., 2010). It has ny elements in the spatial dimension and k elements in the spectral dimension, thus the total detectors are N = K x ny. The wavelength bands for electromagnetic spectrum normally used in spectral imaging basically in three bands which include infrared radiation (IR) which has a wavelength of between 760 nm and 1700nm, visible light (VIS) that range between 400 to 760nm and ultraviolet radiation with wavelength of between 200 and 400nm like as shown in the figure below. Multispectral imaging using infrared Infrared sensors were developed based on focal plane technology. It samples wide range of electromagnetic spectrum that extends from visible band between 0.4μm and 0.7μm to short wave infrared approximately 2.5μm in a number of narrow bands approximately 0.1 nm wide. Most of the infrared sensors operate within SWIR and VNIR bands, utilizing light form the sun for detection and identification of materials depending on their reflection spectra (Dyer, Verri & Cupitt, 2013). The output of infrared or hyperspectral imaging is a number of images over a spectral band, called a cube, which has two spatial dimensions and the spectral dimension is the wavelength as shown in the figure below. The value of the radiant energy is recorded for each pixel data point) for all the wavelength of the sample such that a spectrum for every data point in the image is obtained. An infrared camera can be fitted on quadrocopter to provide a wider coverage range in gold mines. Thermal infrared radiation is used detect the materials that produces more heat compared to the surrounding as they decompose. The sensors use prisms and diffraction grating with 2D or linear detectors arrays to sample the data in a number of spatial bands. The movement of the sensors is along track dimension. To produce such an image in a less expensive design, the camera is fitted with filters to provide flexible and convenient result (Dyer, Verri & Cupitt, 2013). Data representation The way in which data obtained from spectral imaging mainly depends on the data dimensionality. Considering the product of spectral bands times the spatial pixels, this is typically 3-D resolution cells, to the standard for sensor complexity in moving from infrared sensor with narrow spectral bands through reflection of spatial pixels by a factor, but keeping the field of view constant, one dimension spatial reduction compensates for the increase in spectral samples and minimize the required diameter, thus reducing the cost. Although the total number 3-D resolution is preserved, the content of the information is not preserved (Slonecker et al., 2010). Multispectral imaging using visible light Remote sensing involves non-contact monitoring methods which measure the interaction in the energy matter to determine the features of the surface. This technique is mainly associated with overhead imaging like satellite imagery and aerial photography which can record reflected solar energy as a portion of electromagnetic spectrum between 400 to 200nm wavelengths (Moon, Whateley & Evans, 2009). The main application of remote sensing data in gold mining is to identify visual interpretation of land morphological features such as effect of mining on land and natural environment, waste disposal, and transport. Remote sensing has been widely used for the exploration of minerals. Even though gold cannot be detected directly using remote sensing technique, the presence of some minerals like clay and iron oxides are used as indicators to identify hydrothermal region that is related to gold occurrences (Dyer, Verri & Cupitt, 2013). Aerial images have been used for detection analysis of the presence of landfills, and hazardous waste (Moon, Whateley & Evans, 2009). Generally, the images have got enough resolution for detection of small features that can be used for comparative analysis. If the site is monitored over a significant period of time, it can be used to assess the environmental impact associated with mining activities, and also in evaluation of compliance remedial process. The image produced in this technique is interpreted by an analyst and is limited to specific spectral resolutions (Gradus, 2012). The features that can be detected from aerial photography include the impact of mining on vegetation, subsequent use of land after the closure of landfills, local ground movement, and drainage routes (Gradus, 2012). Generally, aerial imagery is a simple and straight forward method because it involves wavelengths that can easily be understood and map like qualities of the images. Visible images include grayscale images or coloured images that falls in the visible part of the electromagnetic spectrum. The coloured infrared image include the near infrared wavelengths between 700 and 900 nm can be used to detect effect of mining on the vegetation (Dyer, Verri & Cupitt, 2013). How to get the ultraviolet illumination Although sensitivity of the camera sensors range between 200 and 1100 nm, it can be attenuated to 350 to 1100 nm by modifying the camera using quartz glass lenses, before filtering the reflection. The resulting image is UV-reflected within the range of 350 nm to 400 nm. On the other hand, the sensitivity of the camera reduce significantly (Murphy & Davidson, 2013). In order to obtain UV reflected image, the camera is arranged as shown below. (Dyer, Verri & Cupitt, 2013; Murphy & Davidson, 2013) UV radiation and the camera placed in positioned shown and UV bandpass filter such as DUG 11X is placed in front of the camera. UV radiation is then turned on and allowed to warm up for 15 minutes. The subject and the reference standards are illuminated before capturing the image. The quality of the image obtained depends on the aperture setting in the center of the lens. The sharpness of the image can be varied depending on the dimension and the distance of the subject. Large aperture settings results in less sensitive sensors to UV radiation (Murphy & Davidson, 2013). Consequently, shutter speed will be quite long. A multispectral image consists of three components which include: Radiation that is produced by the source and travel towards the subject The subject which interact with the radiation from the source The reflected radiation The radiation lies within the three portion of the spectrum. These include the ultraviolet radiation, visible radiation and infrared radiation. The degree in which radiation penetrate into the object is dictated by the wavelength and the material absorbance of the object. The radiation with longer wavelength penetrates more into the object. The radiation reaching the object can be absorbed, reflected or absorbed and re-emitted in form of luminescence with long wavelength. Images produced can either be reflected radiation images or photo induced luminescence or emitted radiation (Siesler, 2002). Reflected images The most common type of reflected images is visible reflected images. It records the reflected light within the visible portion of the spectrum, 400 to 700 nm from the illuminated object when visible light strikes it (Ciancio & Mukerji, 2010). Infrared reflected image provide a record of the reflected radiation within the infrared portion of the spectrum, 700 to 1100 nm, from the object that is illuminated with infrared radiation. The image produced reveals the concealed features. This is due to the fact that infrared radiation has got high penetration and thus some materials like colorants and organic binders are transparent to this kind of radiation (Ciancio & Mukerji, 2010). Ultraviolet reflected images record the reflected radiation within the ultraviolet range, 200 to 400 nm, from ultraviolet radiation illuminated object. The images produced can provide information about the superficial distribution of materials like coating, since this type of radiation can easily be absorbed at the surface. A range of images obtained using multispectral imaging methods described. The images are complementary to each other since each spectral band produce a different components of the same scene to produce a multispectral image set that can be interpreted to provide more holistic scene. The type of interpretation varies depending of the object type (Grattan & Meggitt, 1995; Ciancio & Mukerji, 2010). Image analysis Cameras are design to have built-in adjustments used to vary sharpness, contrast, gain and brightness. However, these adjustments may result in erroneous interpretation of information in the multispectral image. In the Quadracopter, automatic adjustments are suitable. How to produce and receive UV radiation images The acquisition of multispectral images is produced from the combination of different components such as radiation source, filters, detectors and a set of standards used in post processing. A radiation source produces radiation to the subject and also provides means to induce luminescence in the material. Filters allow radiation transmittance within the wavelength range being studied, excluding the unwanted range. A detector consists of digital camera with silicon sensor that has been modified through the removal of infrared blocking filters (Dyer, Verri & Cupitt, 2013). The equipment components required for imaging are shown in the table below. The subject has different combination of filters and radiation source is modified to enable study of a given spectral windows. The recommendation is for the camera characteristics, are made for all methods used. Image technique Radiation source Detector Filters in front of detectors VIS images IR or visible light sources such as LEDS, and tungsten –halogen Digital camera Visible bandpass filter like IDAS_UIBAR Infrared reflected image IR or visible light sources such as LEDS, and tungsten –halogen Digital camera that has been modified by removing infrared blocking filters Filters that blocks visible/UV light UV reflected image UV radiation source like UV LEDS Digital camera UV band bass filter such as DUG 11X UV induced visible luminescence image UV radiation source like UV LEDS Digital camera Filters that can block such as Schott KV418 Visible induced visible infrared luminescence image Visible light like LED lighting Available in digital camera devices. It can be modified by removing infrared blocking filter UV blocking filter like Schott RG830 Visible induced visible luminescence image Visible light like LED source Available in digital camera devices Blue blocking filter and visible light filters like IDAS-UIBAR (Dyer, Verri & Cupitt, 2013) Filters for infrared imaging will block visible and UV radiation that lie within the range 200 nm and 700 nm. These filters allow IR radiation to be collected from 700 nm to 11100 nm. Different materials have different reflective properties within IR range. For example filters such as Schott RG830 cuts on at 830 nm. The sensitivity of this sensor material drops off at large wavelength range (Middleton & Uprichard, 2008). UV reflected images Since silicon based sensors are less sensitive to UV radiation, but more sensitive to IR and visible radiation. Thus IR and visible radiation are blocked, and interferential excitation filter like Schott DUG 11X is used for this purpose. It can be placed in front of radiation source so as to reduce transmittance of IR and violet blue radiation. To obtain UV reflected image, IR and visible radiation of between 400 and 1100 nm are blocked (Middleton & Uprichard, 2008). Output distribution of UV light is shown below. (Dyer, Verri & Cupitt, 2013) The figure shows the wavelengths to be filtered when acquiring UV reflected image. Quartz lens is more appropriate as it makes the acquisition of the images to be at the range of 225 to 400 nm. Radiation and source filters The most appropriate choice for radiation source that can produce sufficient radiation with the required wavelength band as well as efficient filtration of the radiation from the detector is necessary in order to optimize the arrangement set up. The combination of the required filters that can block the radiation from entering the detector while allowing the transmittance of radiation within the desired wavelength. Bandpass filters in front of the camera eliminate the wavelength bands that are not required. Visible reflected images are obtain after the filters allow only the radiation that is visible to the (400 to700 nm) to pass through and blocking any IR or UV radiation. Different bandpass filters can be selected depending on the available filters in different band width within the visible range or it can be on a narrow band focused on given spectral band called monochromatic band pass (Dyer, Verri & Cupitt, 2013). The infrared reflected images are obtained by blocking the visible range and the UV region between 200 to 700nm. The filters allow IR radiation to be acquired between 700 nm and 1100 nm, which are typical for DSLR camera sensors. On the other hand filters are determined based on the type of scene being investigated, because each material has a different reflection features within the IR range. Most of the filters cover between 800 and 1100 nm. Thus filters which have characteristics like those of Schott glass RG830 filter is required as its cuts-on is 830 nm like as shown in the graph below. The graph shows transmittance of visible/UV blocking filter for Schott RG830 (Dyer, Verri & Cupitt, 2013) Conclusion The work presented her shows how multispectral device mounted on Quadracopter can provide substantial benefits to remote sensing and monitoring of gold mine. The device produce detailed, multiple images that provides unparalleled detection and monitoring of the gold mine as well as its effect on the land and natural environment. The images can enhance visual inspection for indication of the presence of Gold or the damage that results from Gold mining. The system is also has the benefit of providing quantification of the inspection results and digital record that can be used for comparison purposes and determination of trend. An important additional advantage for this technology describe here is that the system presents a view beyond the visual images and provide information that may not be possible with physical inspection. It has also been demonstrated that by integrating visible images with infrared and ultraviolet images, more information can be obtained in a more efficient and effective manner compared with the former methods. References Dyer J., Verri G., & Cupitt J., October 2013. Multispectral Imaging in Reflectance and Photo-induced Luminescence modes: A User Manual, The British Museum Grattan, K. T. V., & Meggitt, B. T. (1995). Optical Fiber Sensor Technology. Dordrecht: Springer Netherlands. Gabr, S., Ghulam, A., & Kusky, T. (October 01, 2010). Detecting areas of high-potential gold mineralization using ASTER data. Ore Geology Reviews, 38, 59-69. Middleton, A., Uprichard, K., & British Museum. (2008). The Nebamun wall paintings: Conservation, scientific analysis and display at the British Museum. London: Archetype Publications. Ciancio, A., & Mukerji, K. G. (2010). Integrated management of arthropod pests and insect borne diseases. Dordrecht: Springer. Gradus, Y. (2012). Desert Development: Man and Technology in Sparselands. Dordrecht: Springer Netherlands. Moon, C., Whateley, M., & Evans, A. M. (2009). Introduction to Mineral Exploration. Chichester: John Wiley & Sons. Murphy, D. B., & Davidson, M. W. (2013). Fundamentals of light microscopy and electronic imaging. Hoboken, N.J: Wiley-Blackwell. Siesler, H. W., & Wiley InterScience (Online service). (2002). Near-infrared spectroscopy: Principles, instruments, applications. Weinheim: Wiley-VCH. Slonecker T., Gary B. Fisher, Danielle P. Aiello, & Barry Haack. (2010). Visible and Infrared Remote Imaging of Hazardous Waste: A Review. Molecular Diversity Preservation International. Read More

This report focus on the mapping high potential zones for gold which can be obtained using modified camera used in the quadcopter, which relies on silicon based sensors that covers between 350nm and 1700nm. The cameras that can record infrared radiation with radiation of between 700 and 1700 nm are based on InGaAs sensors, but this technology is often expensive. For this reason, the performance of the existing technologies used in image spectra and gold extraction in n-dimension spectral space is explored (Dyer, Verri & Cupitt, 2013).

Although the measurements wavelength reveals more information about the object, the image obtained may not be interpreted easily. It requires imagery processing to remove all the important information in the spectral bands. Multispectral imaging Every object on the surface of the earth can reflect light in unique patterns; depending on the manner in which light of various wavelengths are absorbed or reflected from the each object. The reflected light is filtered to form unique wavelengths of the electromagnetic spectrum images for different materials.

Multispectral imaging utilizes wavelength band in and outside the visible band of the spectrum. Sampling The collection of spectral image information involves operations that include radiometric, spectral and spatial data imaging. Spectral sampling can be realized through decomposition of radiance received from the spatial pixels into different wavebands that varies depending on their resolution. These bands may overlap depending on the type of sensor. A colored image consists of blue, red and green bands in which the spectral bands do not overlap.

Digital data is obtained from the conversion from analog to digital. The data obtained is three dimensional spectral cubes. The figure below shows that way in which scan lines can be stacked to produce a 3-dimension spectral data cube with information in x, y and z dimensions (Slonecker et al., 2010). It has ny elements in the spatial dimension and k elements in the spectral dimension, thus the total detectors are N = K x ny. The wavelength bands for electromagnetic spectrum normally used in spectral imaging basically in three bands which include infrared radiation (IR) which has a wavelength of between 760 nm and 1700nm, visible light (VIS) that range between 400 to 760nm and ultraviolet radiation with wavelength of between 200 and 400nm like as shown in the figure below.

Multispectral imaging using infrared Infrared sensors were developed based on focal plane technology. It samples wide range of electromagnetic spectrum that extends from visible band between 0.4μm and 0.7μm to short wave infrared approximately 2.5μm in a number of narrow bands approximately 0.1 nm wide. Most of the infrared sensors operate within SWIR and VNIR bands, utilizing light form the sun for detection and identification of materials depending on their reflection spectra (Dyer, Verri & Cupitt, 2013).

The output of infrared or hyperspectral imaging is a number of images over a spectral band, called a cube, which has two spatial dimensions and the spectral dimension is the wavelength as shown in the figure below. The value of the radiant energy is recorded for each pixel data point) for all the wavelength of the sample such that a spectrum for every data point in the image is obtained. An infrared camera can be fitted on quadrocopter to provide a wider coverage range in gold mines. Thermal infrared radiation is used detect the materials that produces more heat compared to the surrounding as they decompose.

The sensors use prisms and diffraction grating with 2D or linear detectors arrays to sample the data in a number of spatial bands. The movement of the sensors is along track dimension. To produce such an image in a less expensive design, the camera is fitted with filters to provide flexible and convenient result (Dyer, Verri & Cupitt, 2013). Data representation The way in which data obtained from spectral imaging mainly depends on the data dimensionality.

Read More