Cluster average shortcomings

[Wes Hand]

I have used trimble access in the past and more recently have watched videos on Carslon GNSS averaging. This is the way point averaging with GNSS should be. Its slick and I have been hoping for a long time that we would come out with something similar. I want to be able to look at the proposed average and see the residuals of each individual point and manually turn off the horizontal or vertical component for each individual point. Although cluster averaging is useful, the only way I know how to cluster average is by setting a predetemined tolerance. Lets say I have mine set to 0.12'. I get 4 shots and 3 are within 0.05' and the 4th is at 0.115. I would like to deselect the 4th shot being somewhat of an outlier from the average solution but if I select cluster average it will put that point with the largest error in there and possibly add some slop into my final positon.

Furthermore, I hope to see us have the ability to shoot the same point again and again with the same number. Other softwares allow it. The 4000.1, 4000.2, 4000.3 gets old quick. If I had never used trimble access I wouldnt know any better and its hard to step backwards when you have seen a better workflow.

Does anyone have a different workflow for averaging that Im unaware of that works better?

 

[Matt Johnson]

Have you tried CoGo Clustering? It creates an average all of points between P1 and Pn that are within a specified Distance Threshold of the average.

 

[Wes Hand]

Hey Matt...thanks for taking my call the other day on the other matter. Much appreciated.

I will try this....ok I see that I had to specify exactly the point range. I just tried it with a point located a total of 5 times as M17.1 to M17.5 and I had to be specific with the M17.5 as the end point. I tried to put in M17.10 treating it like a range of possible points from M17.1 forward and it prompted me that M17.10 does not exist which is true. My workflow involves locating a corner a minimum of 2 times first visit and maybe 3 times if dist to last exceeds my tolerance. I will often visit a second time at a different time/day and gather additonal occupations on any corners that were difficult. With these redundant observations, I would need to keep my point list alphabetical so when I do cluster average under cogo I could see how many M17's I located in the project. I personally like chronological but hey I can change. I do like the displayed information much better than the blind cluster average I've been doing under the additional actions in the points screen but what I'm really wanting the software to do is when I locate M17 the second time as M17 and not M17.2 that it prompt me that it exists with the option for additonal options to store another or average and furthermore upon averaging see a window like you displayed above with the option to tap on or off horizontal & vertical components. I might like the horizontal on an observation but the vertical was poop and only keep the horizontal. I dont see the ability to do so but it would be nice.

So do you manually go to cogo/tools/clustering after each shot? If you do average at the first visit to a corner and then come back to locate additonal occupations later will you not have to manually go in and delete the M17 previously averaged or be forced to average to another number? The option to delete an existing averaged point does not apper to be available within the routine as it is grayed out so I will need to be aware if points have already been averaged. Obviously it will tell me M17 already exists if I try to create M17 again...but I would have to exit and go to the points menu and delete M17 and go back and run it again. Currently I put all averaged shots on a page called AVERAGED and I will delete them if I have to reaverage additonal occuptions with the cluster averaging under the points screen. I do so because it will include the first average and its not an actual observation. This is how I do it but its dangerous as hell as I have deleted several "side" shots that I did not intend to be on the averaged page but placed there by accident. This is too much for inexperienced users and still dangerous for experienced users. I will continue to hope it gets easier.

Sincerely appreciate you taking the time to respond. Its what makes this forum great.

P.S. Is there a way to put all averaged points under 1 report showing the residuals? That would be very nice.

 

[Matt Johnson]

So do you manually go to cogo/tools/clustering after each shot?

Any of the CoGo functions can be assigned to the User Defined U1- U4 hardware buttons.

 

[Wes Cole]

I have used trimble access in the past and more recently have watched videos on Carslon GNSS averaging. This is the way point averaging with GNSS should be. Its slick and I have been hoping for a long time that we would come out with something similar. I want to be able to look at the proposed average and see the residuals of each individual point and manually turn off the horizontal or vertical component for each individual point. Although cluster averaging is useful, the only way I know how to cluster average is by setting a predetemined tolerance. Lets say I have mine set to 0.12'. I get 4 shots and 3 are within 0.05' and the 4th is at 0.115. I would like to deselect the 4th shot being somewhat of an outlier from the average solution but if I select cluster average it will put that point with the largest error in there and possibly add some slop into my final positon.

Furthermore, I hope to see us have the ability to shoot the same point again and again with the same number. Other softwares allow it. The 4000.1, 4000.2, 4000.3 gets old quick. If I had never used trimble access I wouldnt know any better and its hard to step backwards when you have seen a better workflow.

Does anyone have a different workflow for averaging that Im unaware of that works better?

This is also a feature I’d like to see implemented. The way Trimble and Carlson handles multiple observations is a much better workflow. Cluster average is great, and didn’t know about this Cogo cluster average feature until now, but I’d prefer the exact method Wes described. I’m hopeful some features will continue to be implemented in Jfield like the good ol days, but given the lack of software updates recently I’m not overly optimistic.