Tracking Software!!What would that means? please ur opinions

[AutoTrackingTeam]

Hi everybody,

Well I'm not an expert in tracking neither a wilderness survival , I'm an information system student and my graduation project happened to be a software that identify the footprint track of an animal on the sand/mud etc.. and I'm so excited about that coz it would give the chance to learn something interesting and new as tracking.

The initial idea of the project will be as the following:
We will start with limited number of spices from our area to test the success of the idea. The software will be like that a web camera or a mobile one capture a photo of the track, then it'll be send to the software in the laptop where a previous algorithms of each track already had been recorded. The computer will do the image processing and get the result of whom the animal is. Of course as a beginning we will focus on 2 or 3 species and later then develop the program to have more choices and features.

What I want to ask you guys is the following::

From your point of view, What would it mean to you as wilderness survival to have such a software? do you think it's useful? and if not , then Why? do you have suggestion to the idea of the project? and in the first place why is it important to track?

Please guys I would appreciate any responses from you. Feel free to say/suggest whatever you want. I really need opinions to document and start the projects.

Thanks for reading this

 

[Toddy]

Firstly, welcome to BcUK
Why not introduce yourself in the welcomes and imtroductions forum and tell us a little about yourself ?

Tracking:
I think you have two things conflated.
One, a "track", i.e. the footprint of an animal, bird or amphibian which can be recognised from previous experience or personal observation.
Two, tracking, as in following, and making leaps of educated and reasoned logic as to why and where the tracks may be going, using sometimes totally ephemeral evidences, such as crushed grass underfoot.
How on Earth a computer algorythm will help with the latter I do not know.

Personally I would not find such a tracking program of any use; observation of and immersion in the natural world is my relaxation, I leave the computer at home when out.

cheers,
Toddy

 

[AutoTrackingTeam]

Thnx for the replay, I know I've should presented myself but I'm so busy and overloaded with work that's why I did not.
anyway, actually the software is mainly meant to be of help to the researchers about environment and animal species.

and yes , computer algorithms are the main thing if we want to make such software coz through them the computer can process the image and compare the results. Anyway, it's computer vision specialists stuff don't bother .

hmmm .. but I thought the program will be helpfull for trackers also. we intend to analyze the animal features (sex, weight, status running/wounded )etc.

anyway, thanks again

 

[Raz]

I'll echo the above sentiments. the nuances would be undetectable to a webcam, and often tracking, as in, following someone/something is about piecing together partial prints, compressions and using intuition, logic and a combination of sensory clues. Because the tracks aren't usually there.

I don't see it working, too many things to take account for. But if it did, I wouldn't market it as wilderness survival as it'd have extremely limited potential. I'd sell it to schools, scientists and alike. If it is truly quantifiable.

Good luck!

 

[AndyW]

Welcome to BCUK.

I can understand the usefulness from a point of view of learning/training when trying to identify tracks but as Toddy has said, this isn't tracking.

Tracking invlolves understanding your quarry:

Looking for signs of feeding
Identifying scat
Identifying tracks
Considering how it moves
Understanding how it feeds​

This is only a start

I would have also thought that it would involve a high price as it would require a laptop/pda type device. I suspect their would be a market for something like this though, more in the educational environment rather than research though.

There would surely be such a huge range of variables involved when you start to talk about size and weight too. Type of ground, how old is the track, weather conditions of past few days.

I'm not trying to be negative (honest!) but I think this would be quite a challenge.

Please keep us up to date though, it's would be interesting to see how this developes.

Andy

 

[AutoTrackingTeam]

Arigato Gozaimusu.. Hontony Arigatto
don't bother reading .. it's Japanese
Thank you very much .. Really thanks.. Well, I'll take the opinios to the Prof, and see what he would say..

 

[SMARTY]

You could start with the basics EG:

Charactoristics of sign
Tracking terminology
Types of tracking
Tracking techniques
Tracking drills

Just a thought

Smarty

 

[FGYT]

I would Track down Ian Maxwell 'Max'
even go on one of his courses
a 2 day basic course should answer a lot of questions
http://www.shadowhawk.co.uk/

or at least buy his pocket book to Tracking in UK etc

ATB

Duncan

 

[Tadpole]

hmmm .. but I thought the program will be helpfull for trackers also. we intend to analyze the animal features (sex, weight, status running/wounded )etc.

anyway, thanks again

I think it is unlikely that the image from a webcam based system would be of much use, maybe using it in conjunction with some kind of topographical laser measurement of the depth of the track and the direction of foot strike, something to measure moisture in the ground, as tracks are deeper in soft moist, the age of the animal being tracked.
To be honest the variables that would need to be known and pre-programmed before you would get anything near the accuracy needed to identity even a limited number of tracks, in a limited situations, in a single known season, are almost endless. The raw processing power needed would comparable to that of a human brain, along with binocular vision capability, task learning, instant data retrieval and comparison. What you would need is a skilled human tracker

 

[Quark]

I can see the potential in id'ing footprints.
I agree to the above, that trackING, as in following, learning about the animal ect, (as explaind bette by others above) might not be for at image processing software (yet) Humans are better for this... But recently there was some fuss around here in south Norway, about some prints in the snow. At first it was thought to be beartracks, and people were ofcorse alarmed, because bears war not thought to be so far south (yet). The tracks was later id'ed to be wolverine, in it self quite exiting, but not in the same way to sheepfarmers...
But I can see somebody coming across footprints, and wanting to know if it's from wolfe, fox, or domistic dog, or prints from lynx/ big fat house cat... It would be easy to take some pic's send it to somebody with the software and get a ID. Especially to sheepfarmers or hunters. Also to science peopel who track wolfe or other animals would benefit from this, especialy if this software can ID individual animals.

/martin

 

[Ian "max" Maxwell]

Well done and thanks for choosing tracking as your project. I admire people with vision and the ability to look outside the box.
I have worked with several software packages over the years and despite the efforts of the designer they have not really worked. There is the Cybertracker, and a system developed by Zoe and Sky Alibhai to read and differentiate rhino tracks and laterly tiger. The biggest problem is that they dont spend enough time tracking to grasp the shortcommings of their software.
Where ever I go in the world to track I come across rangers who have evaluated the systems. The major challenge is that the same track,from the identical animal will vary every time the animal places its foot down. It is influenced by the angle of the ground and the thoughts of the animal. It is a very challenging prospect, all of which may look different.
The liklyhood of getting a good CIST( Confirmed Identifiable Signature track) is unlikley except of bi-peadal walkers. Anything that walks with four feet will most likely partially or completly destroy the track from the front foot
It will never be able to replace a human tracker, mostly because when we track an emmotional relationship is built with the track, the surroundings and an the animal.
However, I have used software in South Africa, developed by Sasol that works with palm computers. It is for birds and plants. You see the bird, and id it off the computer. It matches it with the sound and other very usefull info, like its favoured food, tree, nest and other birds and mammals its ia associated with. It out-performs carrying lots of id books.
Of course the downside will always be batteries.
One of my assistant instructors and fellow Fgasa/Theta guide who is a very experienced tracker has done lots of work on software and I am sure he would be delighted to help you out.I can see good potential for those that dont have the time to learn to track. I find often that researchers are always on the look-out for time saving devices, and devices that create more of a science( able to repeat the same experiment twice and come out with the same result). :swordfigh For me....I love the experience of being out tracking and all that goes with it including the the colourfull characters around the camp fire!(rain, cold, more rain). Pm me and I will send you my book Animal Tracks ID and Techniques.
Remember....Them That Stick with are them that win!
Good luck.
Max

 

[Rhoda]

I am fairly sure that such a software is already being developed as my business partner has been doing some work on identifying individual animal tracks within a species. Basically being able to identify an individual by it's tracks by photographing front and back feet, left and right, and putting the photos into a piece of software which detects the subtle differences within an individual's tracks and then compares that to the differences between the tracks of two seperate individuals. Therefore enabling the tracker to id a specific individual from a single track by photographing it and running it through the software. (please do correct me if I am wrong Angie!)

Surely if this is possible already then it must be possible to id a species through the software also?

I'm not an expert on tracking technology, as Toddy says I prefer tracking the natural way but it may be worth contacting the company working on this software. I think that they are called Wild Track (pm Bearknights to make sure!)

You should also look up Cybertracker as they have a system where you input data to a handheld system. Your idea may have feasible applications in the field if it could be accessed by a handheld device. That way it could replace the traditional field guides that many of us take along to id species from tracks, especially when in a foreign environment.

To answer your question on why it is important to track I think your best bet would be to book yourself on a course and find out more about tracking in general. If you are going to be working on a project involving tracking then you really should make sure that you are knowledgeable about the subject as a whole.

Hope I have been helpful and not just rambled!

 

[bear knights]

ok a bit of a long reply and probably only of interest to Auto tracker so sorry to rest of you that will fall asleep by the second paragraph!!
You're right in that a similar technology is already being developed. Sky and Zoe Alibhai from Wildtrack work with human trackers who have great experience of working with the chosen species in order to help identify unknown tracks and are developing a Footprint Idenitifcation Technique (FIT). However, the inital algorithms that all other track comparisons are made against are built based on known animals. The work that I have been involved with, mainly bear and more recently polar bear, has used local inuit knowledge to I.D. the individual and determine whether or not it is male/ female, same bear, how old etc and as much as I love to promote working with tracks as a scientifically robust tool for monitoring species, many of the best trackers were misidentifying individual tracks -- for instance saying it was male when there were clear juvenile tracks right next to the adult track and apart from the disneyed version, the juveniles would be with the female and not the male, and also when we caught up with the animal, identification could take place that showed the mistakes being made by human trackers. By developing algorithms based on known animals and then making comparisons between them and new track images, an estimate can be made as to whether it is the same individual or not. we also analysed the accuracy of the human trackers identification of their estimates against known animals and in most cases this helped to support their predictions which for many species, i.e. tiger, tracking as a scientific tool is currently being dismissed as unreliable in favour of expensive and invasive practices such as radio collars, ear tags etc.and so we're hoping our work can support the human trackers claims. Incidently, as someone who has spent many hours tracking, I too was making the same initial identification as the Inuit and although on the whole we were correct in our i.d., mistakes were made. the calculations are still being made as to what level of accuracy we as 'trackers' had compared to the analysis by computer.

The importance of all this analysis is so that actual numbers of that species can be determined and give a real indication of whether that species is as risk.

As an indication, weeks and many km of tracks were followed looking at every individual print of a polar bear in order to find 15 decent images that could be used to identify that individual bear.

I definitely think there is a market for these sorts of tools for anyone wanting to monitor a species remotely without invasive practices, and for anyone trying to promote the sustainability of a project and encourage monitoring by local people instead of making it only for those in the science community.

However I'd have to echo some of the previous concerns in that many of the tracks will not be complete, running tracks can give a diffferent impression to walking, wet mud can give a different impression to dry, etc etc. even the old rule that you can draw a cross through a fox track is not infallible as many dog species also walk on their toes and can give similar results and therefore for someone to be able to use your technology effectively, they would need to be able to understand all the nuances of a single track that make it unique to that animal.

As much as I have misgivings for some of the applications you think you might have for this tool I'd love to see how it develops so please do keep us updated and feel free to PM me if you want some further information on what issues I've faced when involved with this sort of thing before. If identification of tracks can be developed to help conserve and monitor a species and be accepted by the scientific community that is only a good thing.

 

[AutoTrackingTeam]

Waw.. Thanks guys really thanks for taking me seriously, you've been so helpful..

I looked up the other projects mentioned above and yes there are some flaws but the do exist
I'll hold up a discussion session with my team using ur responses and other research items I've been collecting. We still in the re-planning vase and we will try to come out with an acceptable goal of the project.

Since, I'm in the Middle East I think the species of choice at the first level will be from there.
So, I guess (and correct me if u think I'm wrong) it's better to focus on one animal at the time being.Study and cover every thing about it, and if that works out. We can re-do the process with other animals.

I think you are right the camera only may won't be enough and we'll see what we can do about it.

by the way if any one had done tracking in the deserts, would you please let me know

I hope you don't feel like I was wasting your time but the principles and goals of tracking are the same as I've been told.

Thanks again.

 

[bushtuckerman]

I am no tracker but I think, as mentioned repeatedly above, that tracking per se is simply too comlicated to replicate by computer, but that doesn't mean it is impossible.

When designing a piece of technology it is normal practice to look at the natural example and try to emulate it, in this case we have looked at human trackers, but of course humans rely far too much of cognitive processes for them to be replicated yet. I think the real natural example is actually other animals which track their quarry by scent.

This is where technology can truely develop. I read somewhere recently about such a tech and googled it:

http://www.engadget.com/2007/01/27/sniffer-algorithm-leads-robots-to-faint-faraway-scents/

This may be too complex for your project, but maybe it's exactly what you're looking for, so have a look. It's certainly cutting-...nay, bleeding-edge.

The point is, you can only use what computers are good at and then work around the problem. As for webcams, these are far too low a resolution to be of any use. If you want imaging advice then you should research SLR cameras, which use tried and tested algorthms to determine how to expose (meter) for a certain scene. Nikon are the furthest in this field.

 

[FGYT]

i think re reading the original post
hes not trying to make a tracking Computer

but if you find a good print to be able to scan and ID the animal either species or even individual out of a type ie animal 123 of herd/ pride XYZ

still a difficult job but not as much as actually tracking

bit like i do at the moment

Find a Good full print in mud
get out the book and look through till i find a silloutte that matches and try comparing to anything thats similer then read up on other indicators to decide

you could do this by question and answer (which is how you do it in your head)

the imaging could be difficult but it depends what your after and if the image will quickly answer several questions at once ie be quicker than Q&A on its own
i would assume you would have additional parts to the software for research including a GPS stamp for downloading int oa database etc

ATB

Duncan

 

[bushtuckerman]

i think re reading the original post
hes not trying to make a tracking Computer

I got the same impression.

 

[Ian "max" Maxwell]

Looking good,
lots of info flowing on this one.
Several Advanced Shadowhawk trackers have completed their pre-desert and desert tracking specialities and I am sure they would be delighted to offer their help.
Max

 

[SMARTY]

Only "two legged" desert tracking experience im afraid, but willing to help if I can.

 

[AutoTrackingTeam]

Well, thanks for every thing.
I'll be back in 1 to 2 days with more info. and updates. So, I can give a clearer idea and objectives,I'm busy with Exams now. But you were a great help.
will be back.