Lab 10: Radar Remote Sensing

Goal
The goal of this lab is to understand how to perform basic preprocessing and processing of radar images. Specifically noise reduction through speckle filtering, spectral and spatial enhancement, multi-sensor fusion, texture analysis, polarimetric processing, and slant-range to ground-range conversion. When this lab is completed, the basic knowledge of processing on any type of synthetic aperture radar image will be gained.
Methods

Speckle Supression
In this lab radar imagery is analyzed and processed through Erdas Imagine, and ENVI. Noise reduction methods for preprocessing of radar imagery was performed through speckle filtering, edge enhancement and image enhancement. To perform speckle filtering, Radar speckle suppression tool is used. this removes common salt and pepper effect. this is done by utilizing a calculated coefficient of variation (Figure 1).

Figure 1: coefficient of variation is 0.274552.

This variable is then used to run the Speckle Suppression Function in multiple iterations. Multiple iterations are performed to make the image more clear, but it is not necessary to perform multiple, and they can be performed utilizing different filters. The parameters used in this lab are shown in Table 1.

Table 1: In the case of this lab a decreasing coefficient of variation was used, and the window size was increased every pass.

when determining how many passes to perform, the histograms of each image were analyzed. histograms that have spikes present indicate the need for a speckle reduction. In Figure 2 the histogram for the original image, and all 3 despeckle passes are shown.

Figure 2: The top left is the original image, moving to the top right we see slight improvement, but the second and third pass begin to really show improvement, by showing the low portion of the histogram, by removing noise it becomes more visible.

Edge Enhancement

Next edge enhancement was performed over the same lake bed image used for speckle reduction. Two images are created, one where edge enhancement is performed before a Speckle Suppression, and one after a Speckle Suppression. This is done through the use of the Edge Enhancement tool in Erdas. Additionally to see if Speckle Suppression works better before or after an edge enhancement, the non despeckled image with edge enhancement was passed through the Speckle Suppression, but this was not found to be effective. 

Image Enhancement

Next the Image enhancement was performed, multiple options for image enhancement are available in Erdas; Wallis Adaptive Filter, Sensor Merge, Texture Analysis, and Brightness Adjustment. The Wallis Adaptive Filter was used on radar imagery of a glacier.

Sensor Merge

Then a sensor merge between TM and radar imagery over a cloud covered lake was performed. Utilizing IHS principal component, Intensity was used which rescales the greyscale image to the numerical range of the Intensity (I) and substitutes it for I. The multiplicative technique remaps the greyscale image to a 0-1 range. each band then is multiplied sequentially by the remapped greyscale image. 

Texture Analysis

The texture Analysis function is performed in Erdas over an agricultural sub scene from Flevoland, Holland. The texture analysis is performed using imagery from ERS-1 satellite in C-band. 

Brightness Adjustment

This process adjusts pixel digital number values so each constant range has the same average. the brightness becomes even across the image. the input necessary for this process is weather or not the lines of constant range are stored in rows or columns. 

Polarimetric SAR Processing and Analysis

this portion of the lab begen to use ENVI. The imagery used is over death valley, including Stove Pipe Wells, an active sand dune sight. First the image was synthesized, this allows the analyst to receive any polarization combination you want. Three different stretching methods were analyzed Gaussian, Linear, and Squareroot. 

Slant-to-Ground Range Transformation

In this portion the systematic geometric distortion of Slant range radar is the focus. Images appear compressed and stretched due to the changing incident angle. This process takes the slant and eliminates this geometric distortion.

Results 

Speckle Supression

The speckle suppression performed with three passes shows improvement over all the passes, the spikes in the Histograms shown above in Figure 2 being removed are seen in the imagery (Figure 3), by noticing the smoother looking image.

Figure 3: Top Left is the original image, with the bottom right image being the third pass, and a lot of the salt and pepper effect is gone, it looks smoother compared to previous passes.

Edge Enhancement

The edge enhancement allows for greater variability to be seen on ridges, and edges of features (Figure 4). The edge enhancement shows more defined water ripple marks in the lake bed.

Figure 4: This shows a lake bed with defined edges in the image on the right.

Image Enhancement

This portion looked at a glacier, the image enhancement allows to see different features on the glacier more clearly, the image on the right of Figure 5 seems to have more defined features.

Figure 5: Radar Imagery of glacier being enhanced through Wallis Adaptive Filter.

Sensor Merge

In this section TM satellite imagery and radar imagery are combined over an area with a lake covered in clouds. By combining the two, the clouds are removed from radar, which allows the surface feature (Lake) to be seen better (Figure 6).

Figure 6: TM image on left, combined with radar to produce image on right.

Texture Analysis

In the texture analysis portion, agricultural areas in Holland are observed through radar, the texture analysis is perforrmed showing ground surface texture, so crop texture, or soil texture.

Figure 7: Original image on left, texture analysis on right. 

Brightness Adjustment

Brightness adjustment was performed 4 times, and slightly brighter images are produced, but it is difficult to see (Figure 8).

Figure 8: Brightest image should be on the bottom right, but it is difficult to see.

Polarimetric SAR Processing and Analysis

Polametric SAR processing is shown, from each polarization in Figure 9.

Figure 9: Left to right:  HH, VV, HV, TP.

Slant-to-Ground Range Transformation

Slant to ground transformation adjusts and corrects geometric distortion Figure 10. 

Figure 10: geometrically corrected image is shown on the right.

Sources

Erdas Imagine, 2016. 

ENVI, 2015.