I need a good online course on lease squares traverse analysis. Proper field techniques, analysis of the reports, etc.
I don't care if it's accredited for credits...I have plenty...just need to learn a bit more about it.
Any leads?
Tom
Howdy,
For rigorous treatment professors Ghilani and Leick have on-line courses at:
http://surveying.wb.psu.edu/psu-surv/webcorse.htm (Professor Ghilani's site at PSU)
http://www.onlinegnss.com/Courses/courses.html (Professor Leick of Univ of Maine). I understand he has retired. I do not know of his plans for his internet courses.
There are some tutorials associated with specific software packages that are good as well. Do a search for Columbus, Star-Net, TRIMNET, and others. I believe someone on this site posted a link to the digital version of the TRIMNET manual. The NGS software package ADJUST is available free and comes with sample data files as well as guidelines for its use. There is even a NOAA Technical Memorandum TM47 that is quite interesting.
I always hated being asked to analyze the output of an adjustment from a package new to me. There is not a lot of uniformity in the way data is presented.
HTH,
DMM
>not a lot of uniformity in the way data is presented
The free 10-point demo mode of Star*Net is good, but I noticed differences between its reports and Wolf & Ghilani textbook examples. One of the big differences is under what conditions they chose to report a priori vs a posteriori statistics. If the overall fit is better or worse than predicted by the standard errors you give it, then the two may not agree.
> For rigorous treatment professors Ghilani and Leick have on-line courses at:
>
>> http://surveying.wb.psu.edu/psu-surv/webcorse.htm (Professor Ghilani's site at PSU)
ghilani's book, ADJUSTMENT COMPUTATIONS (spatial data analysis) is an EXCELLENT read. book is a little over $100, with sections of probability, statistics, testing, error propagation, etc. the companion CD has a good least squares program. the data files are not as easy to write compared to others, but for that price, it's the place to start
> I need a good online course on lease squares traverse analysis. Proper field techniques, analysis of the reports, etc.
>
> I don't care if it's accredited for credits...I have plenty...just need to learn a bit more about it.
>
> Any leads?
>
> Tom
For an introduction and foundation for understanding the basics, there is this 4 hour video I did on Understanding Least Squares and Positional Tolerance.
Larry P
Learning takes many forms. I am not aware which is best or which you need. But sometimes it just takes "someone to ask."
I have taught least squares for surveyors at the college level since 1980 and have used a number of the books suggested by others. I don't have all the answers but I am pretty good at asking questions. If you want help with an incidental item or want a different perspective on an issue, let me know.
I can be reached through the Global COGO web site.
Before you can jump into that, you need a good knowledge of Least Squares in a pure sense, so take a stats class first.
Once you do that, then you're realize that the traverse is an iterative process and the lower your iterations are, the better traverse you have.
For a single loop, compass and roll on. If, you have lots of GPS hubs, cross ties you can evaluate, etc., then LS may help you. Otherwise, if you still run the same old loop, it doesn't help, at all. The main thing you need is extra degrees of freedom, of in this sense, extra measurements. That can't be gleaned from the simple loop, but only from GPS ties or triangulation ties.
You need to know stats first, or you'll never get it.
:good: Agreed. The prereq's to the least squares course at GBCNV are Stats and Calc I
I attended a seminar presented by Ghilani at the ACSM Conference in 2004, when it was in Nashville. It was a good class, very intense, and some parts were way over my head.
Kris and David are correct. A good class in Statistics is a must. That was the part that lost me.
Larry,
I have that class in my wishlist on your website. When things pick up a little, I will work that into the budget, along with a few other books I have in there as well.
Take care my friend,
Jimmy
> You need to know stats first, or you'll never get it.
I agree with that statement, to a degree, but I don't think it necessary to take a college class in statistics. A good reading of "Statistics for Dummies", or the like, is in order. You need some grasp on concepts but the program does the math for you.
