Congratulations on nearly 9 years of successful operation. However, the dynamics and interests of citizen volunteers has evolved and grown less over the years. I would like to submit a few suggestions:
1. Given the very small pool of volunteers, I think the Image Entries pick list should be swapped out much sooner, say after 2 weeks. Even if only one person has submitted a proposed finding, keeping images in the pick list for a month or more just hoping for more responses just means alot of other images are never going to be found.
2. I am not sure that the points system is all that useful anymore. It *is* however nice to know that others are actively volunteering. Maybe just a "quantity submitted" each week is all that is needed. No one is "competing" for high scores anymore (if ever).
3. As you know, most of the photos are in "sets" at the same relative location. We could easily geo-locate those 5 to 10 additional nearby photos so a much larger set of photos in the library are marked as "Cataloged with Center Point". As was your very first feedback on 23 July 2013, let us choose additional photos to geo-locate or at least attempt to geo-locate those +/- 10 before and after the photo in the pick list in filename order.
4. I use custom-made web tools that show me an entire series of photos before and after the "picked" image in your list, Google Earth for easy general geo-locating, and frequently use the existing "Search Photos" capability with ranged LAT/LON from the sub-orbital point. But most volunteers are just not aware how much easier those tools can make the task of geo-locating an image. Mention and recommend the use of Google Earth in your tutorial and FAQs. Also adding two new buttons on the primary Image Detection geo-locating page would vastly help volunteers: "Show Search Results" and "Show Addtl Images".
The "Show Search Results" would launch your existing "Search Photos" capability's results using "widened" Lat/Lon from the sub-orbital points with just "Include Cataloged with Center Point" checked (plus Daytime & Nighttime Images, plus all ISS missions selected). The "Show Addtl Images" would show +/- 10 images before and after the picked image's filename. Many times there are "context photos" before or after the images in your pick list that are far easier to geo-locate and lead directly to the image on the pick list. Do please provide the ability to click on a photo and bring up a *full-sized* copy of the image in a separate browser window is essential.
5. Much like the "Cities at Night" project, if the Image Detectives system had an additional "task" allowing volunteers to review and agree/disagree with submitted findings would help YOUR review process significantly, and let volunteers that may get frustrated with geo-locating images to still help by reviewing proposed submitted findings.
6. It likely would also help your internal review process (or that of #5 above) if the main Image Detective screen had a blank text field allowing volunteers to document just HOW a particularly difficult image was found. Many times I find a very difficult image by tracing through 10 sometimes 20 prior or next images. Just being able to supply you'all with the prior image or images that lead me to your picked image would help reviewers significantly in confirming a difficult-to-find geo-location since they could also "trace that path".
Again congratulation for a long-term citizen volunteer project, and thanks for any consideration of these suggestions.
The two banners at the top of the new primary image matching web page take up way too much of the screen real estate! Matching was much easier without this dead space at the top of the Google Earth display.
I've been a volunteer working on the CitiesAtNight.org project, and have personally identified and geo-referenced over 400 of the NASA images thus far, and I know from their database that there are between 1,000 and 2,000 NASA images that are potentially geo-located. I have two questions:
1. Does the staff at NASA working on the NASA Image Gateway project and the Image Detective planning on pulling and incorporating the information that CitiesAtNight has vetted and will be vetting as the work progresses? It would seem an invaluable resource that would greatly amplify the Image Detective work.
2. If the NASA staff had not planned to gather and incorporate those identifications from CitiesAtNight, then rather than just having the small limited pool of pick-a-photo the current Image Detective offers, can an Image Detective "bug" be added to every NASA Image Gateway page, such that an Image Detective can click that bug/web link, and have that image loaded into the Image Detective application? It would be just a small extra step on my and other volunteers part to call up and post the identification into the Image Detective app in addition to doing it in the CitiesAtNight app. It seems so unnecessarily limited to only offer 96 images to choose from at a time in the Image Detective app, given the 100,000s of images available. I personally geolocate between 10 and 20 NASA images per day just in my spare time on CitiesAtNight.
A growing quantity of accurate identifications are being made for the NASA images, and it seems a tremendous loss for NASA to not tap what has already been identified. I personally do not care about points and top-scores and the like; I just want the identifications already made and continually being made to be as widely available as possible in NASA's own database.
Is it possible to add a "I can't quite locate it button"? 25 Aug 2014 MoeZen raised the "uncle" button question and the NASA response was that the frame he is viewing had a wrong location. I am having problems with several photographs, I am being pointed at the correct area I am sure, but I just can't make the find! Examples: ISS005-E-14894 & ISS036-E-36037 in my queue. They just sit there and I can't find them! Considering ISS036-E-36-37, I have already identified 20 or more pics from that region, but I just can't locate this one, yet it hangs in on my queue. How about creating a "I can't get it button" that allows us to annotate with a comment about where we believe it is and allow users to get to it a take a shot. This would be very effective collaboration! ...and I agree with MoeZen, there should be higher point value for one of these "difficult" pics!
Some of the photos are beautiful fractals (most difficult to locate them anyway because of the self-similarity) so my idea is to annotate these photos by a "fractal" label and later it would be possible to create "Geological fractals on Earth" collection of photos or database that might be interesting for the public and for the scientists too.
Thank you for adding pages to the set of images we can select. I was often running out with only five. And I notice you are now sorting the images by point value (I'm assuming the higher point images are more important to you), which is also nice. There are two further refinements that would be nice to have. 1) List the images from highest point value to lowest point value (the opposite of how they are listed now). That way it's easier to start with the most important ones. 2) Add a field so that we can jump to a particular page rather than having to use the left/right arrows. This will save us a fair amount of time. Thanks.
I think accuracy could be increased if it was possible to move the (center) point in the photos. Lining up the center points can be difficult. If there was a way to move the point in the photo, then specific geographic/geologic features could be used for pinpoint placement. In some instances the placement could be withing a few meters of the true position.
Stargazer: completely separating image location from annotation of the visible features. This looks like a really good sophistication for our project in the longer term. Different sides of the brain!
MoeZen: The image ISS036-E-16046 seems to have a wrong camera time in original metadata, a problem that occasionally occurs. But the image was subsequently cataloged (http://eol.jsc.nasa.gov/scripts/sseop/photo.pl?mission=ISS036&roll=E&frame=16046&QueryResultsFile=140908989470682.tsv ) and shows up as Memphis, TN airport. Sorry you had a bad experience with this frame!
CMDRSHEPARD: -- It certainly would be helpful to recreate the previous settings only ONE click away -- "do what I just did on the next image" button seems like just the kind of solution that a computer should be able to solve. We'll look into it. -- recurring to the last pick page looked at ought also to be doable.
Suggestion. Since the objective is to catalog the million+ photos it would make sense to completely separate the locating of the image and the annotating of the features visible. I like finding the locations of each image but don't feel qualified to be annotating the geological features that are present. Perhaps two different steps (zones) can be offered for earning points. Let some people find the locations, center points, zoom, rotate, etc. while others could simply be asked to annotate the features of the pictures given the locations. I think there are many people who might offer their time annotating who are not as good at searching. (Find once, annotate many.) Once a location is found it should be grayed for everyone. No sense wasting time having everyone find the same location. All detectives would be encouraged to annotate features. I believe today if a person enters features observed without a good location it may be confusing to the validation team. Only add features to "found" locations. Separate "Pick Lists", one for those needing locating, and a different "Pick List" for annotating. (Just some thoughts.) (Having fun finding these images.)
Would it be possible to have a button that would let us give-up on a photo? For lack of a better term, it could be called the "Uncle" button. Maybe by saying "uncle" the point value could be increased and the photo could either be passed to another user, or put in a public queue where any user would have to chance to identify the location. Photo ISSO36-E-16046 drove me nuts for, sadly, about 4 hours (cumulatively over a few days). I don't give-up on puzzles, but even grid-type searches, of a very large area, didn't work. Now I'm curious!
I don't know if it would be difficult to implement, but it would make things more efficient if there were some optional button to press after submitting an image that would:
1) keep the Google map zoomed in to its current location 2) pull up the next available image.
Bonus points if it would also rotate and zoom the next image to the same degree as the previous (submitted) one.
Astronauts seem to like to snap several images of the same thing in quick succession and it would save a lot of clicking. Currently after submitting an image I choose pick image-> click to the page containing the image I want (say it's page 5) -> click the image -> hide the image -> zoom the map and scroll to the correct location -> show the image -> click the rotate button several times -> resize the image -> choose cloud cover -> type features -> click submit. That's probably 20+ clicks for an image that is nearly identical to the last one.
This would also make my results more standardized. For example, we currently have about 12 images of Greece that are nearly the same. Because they were taken at an angle, they were pretty foreshortened. I had difficulty knowing how much to zoom and rotate because of distortion so the amount of each I did was probably wildly variable between images.
Barring a magic "do what I just did on the next image" button, it would make things easier to be able to type in the amount of rotation we want. If I know I rotated the last image 117 degrees it would save me a few clicks to just enter that number for the next image. Otherwise I need to click +45 +45 +45 (change to 5s) -5 -5 -5 (change to 1s) -1 -1 -1.
A way to enter which of the 5 pages of images we want would also be appreciated. Or have the pick image screen popup to the last page we chose from instead of automatically returning to the first page. Or have the pick image page arrows wrap around so we can get to page 5 from page 1 by clicking the left arrow instead of clicking right arrow 4x.
It's not a huge deal if these things can't be implemented (it's not like we're on production pay or anything), but when faced with 10 or more nearly identical images, it's a little discouraging knowing you'll have to start at zero for each one (200+ clicks). I think people who don't have much time to spare or who are more casual would be put off from starting a bunch of pictures of the same thing or would start them but get bored and quit in the middle.
The rotation of the image could work better. The rotation cuts off the corners of the image because the rotation window does not rotate with the image.
For images that are almost the same, it would be great to have a button that starts the next image in the same location as the last image, to make it quicker.
The Pick List is updated once the images have been identified and submitted for approval by any user. Once the images are identified, new images come in to replace the existing images.
As for seeing the same images in the Pick List: it is very likely that you did not improperly locate the image. We typically have access to review new imagery that our Image Detectives submit the day after you submit them. Therefore, in the mean time, they may show up in your Pick List until we have had the chance to approve them. We have recently been notified of an issue that images a user has already identified do not show up as "grayed out", or unavailable to re-identify. That issue is currently being fixed.
Thank you for all of your feedback. We will plan on creating a PDF version of the tutorial so users may have that asset as well.
tallbob29: Thank you for identifying the issue of images not graying out after you have identified them. We are currently working on the issue, and it should be fixed shortly.
How about adding a way for users to "opt in" to show their email addresses. Or make a true forum board, where registered users can post and respond to other users. I could see this becoming like other forum boards.
I found my way here a few days ago from the CNN article and am really enjoying seeing all of the amazing photos. Can you tell me how often the available images are changed? Is it done for each detective or are they generally available for a set period or until a certain number of people have identified them? The first couple of days it seemed like they turned over daily, but now I feel I'm seeing many I've already completed. I'm not sure if they're still available to me because I didn't do them properly the first time and I should do them again or if I should wait for new ones to appear.
Thanks for making an easy to use interface. The tutorial was a good length (not too long) and straightforward. For some reason I was expecting narration but it's not necessary. A downloadable PDF might be used by some.
You might want to add a tip sheet on how to approach foreshortened images. I find it hard to decide how much to zoom.
It would be nice if you put a post (maybe in the General topic, or in a FAQ topic) that explained what happens to these images once we submit them. It would be interesting to know the remainder of the process that goes on and where they end up.
It would be nice if, when I start a new session, the images that I've submitted in previous sessions are not presented to me on the "Pick an Image" pages. For example, today, when I logged in, about 4 of the 5 pages contained images that I'd submitted yesterday (or the day before), and only about 10 new images.
It would be nice, on the 5 "Pick an Image" pages, if the text for images I've already submitted stayed "grayed out" for my entire session. At present, when I submit images, the text stays grayed out for a while, but then goes back to white. Bottom line: The text color (gray or white) is not a reliable indication of whether I've submitted the image already, or not. So I must keep track separately.
I believe having the ability to see your own score is important. Frankly, I have submitted several items and have not been able to find myself on the score chart. How do I know my submissions are counting for something?
geoDude55: Another good suggestion, but one that might require a little more work and take some time to implement. We will certainly look into it as an alternative, as it does sound very user-friendly.
geoDude55: Thank you for your feedback! We have added the sun azimuth angle to Image Detective for user convenience. This angle is the clockwise angle in degrees from north to the sun, and as geoDude55 mentioned, this should aid in rotating the astronaut image when locating the center point.
geo-positioning the pictures can become easier: let us mark 2 far apart points on the picture, and their corresponding points on the map. Render the picture joining the corresponding points. Extra enhancement: adjust the projection of the map according to the point of view of the spacecraft. Google Earth can do that. Regards, GeoDude55
One suggestion I would have, is to have your "points" visible on what you've earned up to this point. Unless I'm missing it, the only place I can see that you'd find it, is by searching thru a long list of other "detectives" to see where you might be in the list. Just would be nice to see it on the actual Image Detective Screen while you're searching for the next image. :)
Please add to the photograph info the direction of the sun in degrees relative to north and zenith (degrees N/S and E/W relative to zenith is also possible). I just saw clear shadows on a "difficult" to spot picture of mountains. With the direction of the sun, I can rotate the picture to fit the shadows with the direction of the sun.I guess you NASA girls and guys don't have to be told, that the sundirection can be calculated from the date and time the photograph was taken and the spacecraft suborbital point. Regards.
I think it would be interesting to pick or submit a request for our own selected images that we want to locate from the uncatalogued images on the Gateway to Astronaut Photography of Earth website. Does anyone else think they would like to be able to pick some images for themselves to locate?
36 entries
To add content, log in to the Image Detective system after your detective name has been approved.
