High-Level Project Summary
We developed a tool to process the JunoCam raw images to generate images with more details and convey more meaningful information easily discovered by close observations. We have developed image denoising, edge enhancement, color enhancement and low light enhancement methods to process the images.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Jovian images are blur and lack of details in general. Some images are also too dark to see. We process Jovian images by applying image processing methods to make the image details more visible. We also manipulate the image in the HSV color space so that the resultant image is more color vivid, which also reveals more details. For very dark images, we revert the image components, apply haze reduction and then revert the image back. It make the image more brighter and shows more information with human vision.
To reveal more image details, we first covert the image to YUV space. Then we focus on the Y, grayscale image. We need to make the Y image more smooth. Then by subtraction the smoothed image from the original image, we get detailed grayscale image. The detailed grayscale image is then amplified and then added back to the original image. Optionally, we can apply histogram equalization and/or gamma correction to show more details hidden in the dark. By converting the image back to RGB color space, we get the final, brighter, and detail enhanced color image.
To take the image color more vivid, we first covert the image from RGB space to HSV space. Here H stands for Hue, S for saturation and V for brightness. Once we get the S component, we can then amplify it to make the image color more saturated. To show image with more details, we apply the same image enhancement method as before on the V component. When converting back to the RGB space, we get an image with more vivid color and details.
All the image processing methods are implemented with open source script language Octave.
Space Agency Data
Junocam images were used. The images are in general blur and lack of details perceivable by human vision. Some image are too dark to check the details. Some simple but effective image processing techniques can help to enhance the image details and more properly relighted the image.
We expect after image processing, the processed image can reveal more information not noticed before.
Hackathon Journey
We got interested in the topic of image processing because we seen this topic often discussed at the workplace of one of the team members. In performing this, we have obtained a basic understanding of image processing and how smoothing/equalization/gamma correction is done. Various components of the project were completed, a script for video presentation was prepared, details of image processing discussed. Also to demonstrate the success of the image processing software, a special scratch file was created to present the success of making better images of Jovian system. What was discovered that cameras that originally take the pictures see the outside reality differently and operation such as gamma correction are needed to control the brightness. A lot of information was learned, such as new terms, methods, coding rules. Issues in coding such as things not being times correctly and pieces of code not matching. But overall team members were proud of their participation.
We want to thank our host, the Magikid Lab in San Jose, for providing accommodations for the whole event
References
https://www.missionjuno.swri.edu/junocam/processing?featured=1
https://www.missionjuno.swri.edu/junocam/processing?source=junocam
We learn the basics of image processing from Pennsylvania State University website
Tags
#imageprocessing #programming