GE 110 Remote Sensing Lecture 6 PDF

GE 110: REMOTE SENSING Lecture 6: Image Rectification and Restoration ENGR. ARTURO G. CAUBA JR. Instructor I College of Engineering and Geosciences Caraga State University GE 110: Remote Sensing Lecture 6 Outline ▪ Image Rectification and Restoration Concepts...

GE 110: REMOTE SENSING Lecture 6: Image Rectification and Restoration ENGR. ARTURO G. CAUBA JR. Instructor I College of Engineering and Geosciences Caraga State University GE 110: Remote Sensing Lecture 6 Outline ▪ Image Rectification and Restoration Concepts ▪ Geometric Correction ▪ Radiometric Correction ▪ Noise Removal GE 110: Remote Sensing Lecture 6 Expected Outcomes The students would be able to: Learn the concepts behind image rectification and restoration; Identify the various computer-assisted procedures of image rectification and restoration; and Learn how to conduct the computer-assisted procedures through laboratory exercises GE 110: Remote Sensing Lecture 6 IMAGE RECTIFICATION AND RESTORATION CONCEPTS GE 110: Remote Sensing Lecture 6 Image Rectification and Restoration (1) Operations that aim to correct distorted or degraded image data to create a more faithful representation of the original scene. Often termed “Image pre- processing” operations/procedures: They are normally done before further manipulation and analysis of the image data to extract specific information Most of the distortions and degradations are caused by several factors during the image acquisition process.  GE 110: Remote Sensing Lecture 6 Image Rectification and Restoration (1) Typically involves initial processing of raw image data to: Correct for geometric distortions i.e., to ensure that all pixels in the image are correctly geo-referenced “GEOMETRIC makes it possible to conduct accurate point, line and area CORRECTION” measurements in the image Calibrate/correct the data radiometrically: “RADIOMETRIC e.g., to convert DN to absolute radiance values, to correct for CORRECTION” atmospheric effects, to correct for changes in scene illumination Eliminate noise present in the data “NOISE REMOVAL” e.g., to remove stripes, bit errors, etc. GE 110: Remote Sensing Lecture 6 GEOMETRIC CORRECTION GE 110: Remote Sensing Lecture 6 Geometric Correction Why is it needed? – Raw digital images usually contain significant geometric distortions – These distortions make the raw images unusable e.g., they cannot be used directly as a map base without subsequent processing.  GE 110: Remote Sensing Lecture 6 Example of a Distorted Landsat Image  GE 110: Remote Sensing Lecture 6 Example of a Distorted Landsat Image with Road Network GE 110: Remote Sensing Lecture 6 Example of a Geometrically-corrected Landsat Image with Road Network GE 110: Remote Sensing Lecture 6 Some Sources of Geometric Distortions 1. Variations in the altitude, attitude, and velocity of the sensor platform 2. Earth curvature 3. Earth’s eastward rotation 4. Atmospheric refraction 5. Relief displacement By applying geometric correction procedures, the distortions introduced by these factors are compensated so that the corrected image will have the highest practical geometric integrity. GE 110: Remote Sensing Lecture 6 Distortions caused by variations in the altitude, attitude, and velocity of the sensor platform From Lillesand et al., 2015 GE 110: Remote Sensing Lecture 6 Distortions caused by Earth’s curvature The Earth's curvature affects the geometric scale and exerts a type of panoramic effect Commonly observed in images acquired from high altitudes More pronounced at higher latitudes A non-systematic type of distortion From: Gupta, R.P., 2013. Remote Sensing Geology, Springer Science & Business Media. GE 110: Remote Sensing Lecture 6 Distortion caused by the Earth’s eastward rotation The eastward rotation of the Earth during a satellite orbit causes the sweep of scanning systems to cover an area slightly to the west of each previous scan. The resultant imagery is thus skewed across the image. This is known as skew distortion This distortion is systematic Common in imagery obtained from satellite multispectral scanners. From: Gupta, R.P., 2013. Remote Sensing Geology, Springer Science & Business Media. GE 110: Remote Sensing Lecture 6 Distortion caused by atmospheric refraction Serious error in location due to refraction can occur in images formed from energy detected at high altitudes or at acute angles GE 110: Remote Sensing Lecture 6 Distortion caused by relief displacement Relief displacement = displacement in the position of the image of a ground object due to topographic variation (relief) Common phenomenon on all remote sensing data products, particularly those of high-relief terrain From: Gupta, R.P., 2013. Remote Sensing Geology, Springer Science & Business Media. GE 110: Remote Sensing Lecture 6 Relief Displacement Schematic From: Gupta, R.P., 2013. Remote Sensing Geology, Springer Science & Business Media. GE 110: Remote Sensing Lecture 6 Example image with relief displacement (1) Lecture Notes in GE 113: Remote Sensing 19 TOPIC 6. IMAGE RECTIFICATION AND RESTORATION GE 110: Remote Sensing Lecture 6 Example image with relief displacement (2) Lecture Notes in GE 113: Remote Sensing TOPIC 6. IMAGE RECTIFICATION AND RESTORATION 2 0 GE 110: Remote Sensing Lecture 6 Example image with relief displacement (3) Lecture Notes in GE 113: Remote Sensing TOPIC 6. IMAGE RECTIFICATION AND RESTORATION 2 1 GE 110: Remote Sensing Lecture 6 Geometric Correction Process A 2-step procedure: – Correction of systematic distortions (e.g., skew distortion) – Correction of non- systematic/random/unpredictable distortions GE 110: Remote Sensing Lecture 6 Correction of Systematic Distortions Easily corrected by applying formulas – Sources of distortions are mathematically modeled – Example: Skew distortion due to earth’s eastward rotation: – Corrected by deskewing the imagery » Involves offsetting each successive scan line slightly to the west » The skewed- parallelogram appearance of satellite multispectral scanner data is a result of this correction From: Gupta, R.P., 2013. Remote Sensing Geology, Springer Science & Business Media. GE 110: Remote Sensing Lecture 6 Correction of Random Distortions (1) Corrected by analyzing well-distributed ground control points (GCPs) occurring in an image – GCPs are features of known ground location that can be accurately located in an image GE 110: Remote Sensing Lecture 6 Correction of Random Distortions (2) Correction Process: 1. Numerous GCPs are located both in terms of their two image coordinates (column, row numbers) on the distorted image and in terms of their ground coordinates 2. These values are then submitted to a least squares regression analysis to determine the coefficients for two coordinate transformation equations that can be used to interrelate the geometrically correct (map) coordinates and the distorted-image coordinates 3. Once the coefficients for these equations are determined, the distorted image coordinates for any map position can be precisely estimated GE 110: Remote Sensing Lecture 6 Mathematical Notation of the Transformation Equation x = f1(X,Y) y = f2(X,Y) Where: (x,y) = distorted-image coordinates (column, row) (X,Y) = correct (map) coordinates f1, f2 = transformation functions  GE 110: Remote Sensing Lecture 6 Correction of Random Distortions (3) Correction Process (continuation): 4. Using the transformation equations, a process called resampling is used to determine the output (“corrected”) image matrix from the original (“distorted”) image matrix: The coordinates of each element in the undistorted output matrix are transformed to determine their corresponding location in the original input (distorted-image) matrix The intensity value or digital number assigned to a cell in the output matrix is determined based on the pixel values that surround its transformed position in the original input From Lillesand et al., 2015 matrix GE 110: Remote Sensing Lecture 6 Resampling Methods Nearest neighbor – The DN of the transformed (“corrected”) pixel is equal to the DN of its closest (“original/distorted”) pixel – Advantages: simple to implement Avoids the alteration of the original pixel values – Disadvantage: Features in the output image may be offset spatially by up to one- half pixel Causes disjointed (“blocky”) appearance in the output image GE 110: Remote Sensing Lecture 6 Resampling Methods Bilinear interpolation – The DN of the transformed (“corrected”) pixel is equal to distance-weighted average of the 4 nearest pixels – Advantage: Smoother image appearance – Disadvantage: Alters the original DN values Nearest neighbor Bilinear interpolation  GE 110: Remote Sensing Lecture 6 Resampling Methods Bicubic interpolation or cubic convolution – The DN of the transformed (“corrected”) pixel is determined by evaluating the block of 16 pixels in the original image that surrounds the output pixel – Advantage: Smoother image appearance Slightly sharper image than the bilinear interpolation method – Disadvantage: Alters the original DN values GE 110: Remote Sensing Lecture 6 Other Uses of Resampling Methods (aside from geometric correction of images) Used to overlay or register multiple dates of imagery (“image-to-image registration”) Used to register images of differing spatial resolution Used extensively to register image data and other sources of data in GISs. GE 110: Remote Sensing Lecture 6 RADIOMETRIC CORRECTION GE 110: Remote Sensing Lecture 6 Radiometric correction Why is it needed? – The radiance measured by the sensor is influenced by different factors and it must be corrected – Factors affecting radiance: Changes in scene illumination Atmospheric condition Viewing geometry Instrument response characteristics (e.g., how does a sensor record radiance as DNs?)  GE 110: Remote Sensing Lecture 6 Examples of Radiometric Corrections Sun elevation correction – Accounts for the seasonal position of the sun relative to the earth – Image data acquired under different illumination angles are normalized by calculating DN values assuming the sun was at the zenith on each date of sensing GE 110: Remote Sensing Lecture 6 Examples of Radiometric Corrections Earth-sun distance correction – Applied to normalize for the seasonal changes in the distance between the earth and the sun Earth-sun distance = in Astronomical units 1 Astonomical unit = mean distance between the earth and the sun = 149.6 x 106 km – The irradiance from the sun decreases as the square of the earth-sun distance  GE 110: Remote Sensing Lecture 6 Examples of Radiometric Corrections Combined sun elevation and earth-sun distance corrections: Where: E = normalized solar irradiance E0 = solar irradiance at mean earth-sun distance θ0 = sun’s angle from the zenith d = earth-sun distance during the acquisition, in astronomical units GE 110: Remote Sensing Lecture 6 Examples of Radiometric Corrections Atmospheric correction – the large amounts of imagery collected by the satellites are largely contaminated by the effects of atmospheric particles through absorption and scattering of the radiation from the earth surface. – The objective of atmospheric correction is to retrieve the surface reflectance (that characterizes the surface properties) from remotely sensed imagery by removing the atmospheric effects. – Atmospheric correction has been shown to significantly improve the accuracy of image classification  GE 110: Remote Sensing Lecture 6 Examples of Radiometric Corrections Conversion of DNs to absolute radiance – DNs are converted to spectral radiance using the sensor’s radiometric response function – Such conversions are necessary when changes in the absolute reflectance of objects are to be measured over time using different sensors (e.g., the TM on Landsat-5 versus the OLI on Landsat-8). Likewise, such conversions are important in the development of mathematical models that physically relate image radiance or reflectance data to quantitative ground measurements (e.g., water quality measurements). GE 110: Remote Sensing Lecture 6 Example sensor-specific formula relating DN with Spectral Radiance GE 110: Remote Sensing Lecture 6 NOISE REMOVAL GE 110: Remote Sensing Lecture 6 Image Noise Any unwanted disturbance in image data that is due to limitations in the sensing, signal digitization, or data recording process Noise can either degrade or totally mask the true radiometric information content of a digital image GE 110: Remote Sensing Lecture 6 Potential Sources of Noise Periodic drift or malfunction of a detector Electronic interference between sensor components “Hiccups” in the data transmission and recording sequence GE 110: Remote Sensing Lecture 6 Example of Noise in Image Stripes or bands – Appearance of defective lines (e.g., in Landsat MSS data) due to the variations in the response of individual detectors Results to relatively higher or lower values along every sixth line Line drop – A number of adjacent pixels along a line may contain spurious DNs – Caused by data transmission errors Bit errors or salt and pepper noise – Random noise in the image – “spiky” in character – Causes images to have a “salt and pepper” or “snowy appearance” GE 110: Remote Sensing Lecture 6 Example of Striped Image GE 110: Remote Sensing Lecture 6 Example of Image with Line Drops GE 110: Remote Sensing Lecture 6 Example of Image with Bit Errors GE 110: Remote Sensing Lecture 6 Noise Removal Process Usually precedes any subsequent enhancement or classification of the image data Correction/removal depends on the nature of noise: – Systematic (periodic) – Random – Combination of systematic and random noise Noise removal are done through: – Interpolation of the DN values – Application of moving window algorithms GE 110: Remote Sensing Lecture 6 References/Further Reading Lillesand, T. M., Kiefer, R. W., & Chipman, J. W. (2008). Remote Sensing and Image Interpretation 6th Edition. United States of America: John Wiley & Sons, Inc. Online Tutorial: Fundamentals of Remote Sensing – “Pre- processing”. Available at http://www.nrcan.gc.ca/earth- sciences/geomatics/satellite-imagery-air- photos/satellite-imagery- products/educational- resources/9403 GE 110: Remote Sensing Lecture 6 Thank you for listening! ☺☺☺