What new area's of the surveying/spatial data capture industry are you paying attention to because you foresee they may gain market share as time goes on?
Ok coarse there are the obvious ones like aerial lidar, which is X years will probably replace the vast majority of "hand" topos as sensors continually get better, cheaper and smaller and processing gets easier and faster. Or terrestrial/SLAM lidar for all the same reasons.
But are there any you've noticed that aren't getting as much attention? Airborne bathymetry? Scan to BIM with automation via machine learning? Integrating SUE with AR? The creation and implementation of "true" digital twins? Auto drafting/calcing plats via Image recognition & AI?
It certainly depends on your firm's needs and type of work they do. For us, we've seriously been considering the SLAM technology. It seems like game changer to have most of your field work topo acquired in less than a day. The amount of detail and purported accuracy is mind boggling.
While hard surfaces seem to give good scan results, I've seen photos like this where people have suggested using scanners. I very much doubt you would get a good surface model...
We've been using SLAM for a few years now. The data capture is fast but that's really the tip of the iceberg. Marketing for SLAM tends to only compare field acquisition time vs traditional methods and conveniently ignore data extraction time.
It has it's uses but do we try to do site plans with it? We do not. It's reserved jobs that require modelling as a deliverable.
I have a video comparing SLAM to aerial lidar as far as accuracy and veg penetration goes...
Jim, aerial LiDAR would be fine in that situation. I wouldn't hesitate for a second on creating a surface from a flight covering that area. The right person could strip away that vegetation quite easily. Remember, LiDAR sensors are capable of multiple returns. Even cheap sensors (DJI L2) have 5 returns now.
Would it produce a better surface than a total station? Not at the exact spot where the shot was taken but everywhere else, yea. We must not underestimate how important data density is.
I wouldn't walk it with a SLAM unit, but only because I wouldn't want sticks brushing up against my sensors, not because I think it wouldn't penetrate.
Yeah flying lidar with correct signatures one could eliminate the high vegetation. It’s all about the signatures. Once you solve that you can do well.
I was talking to the head of a large county department of public works about using SLAM for environmental surveys. Scan a wooded parcel, build a ground tin, raise it 4.5 feet and cut a section - you've now measured all the trees diameters.
Thanks Lukenz
We tried that James. We ran into 3 issues;
1. The SLAM algorithm getting lost in an environment void of planes and in constant movement (leaves blowing in the wind).
2. IDing the trees from the scan data was difficult depending on tree species.
3. After traversing through the wooded areas, scanning the forest X amount of times until you had a data set void of drift, extracting data from the cloud and IDing trees from the cloud or having to go through imagery to try to ID species, it wasn't a cost effective solution.
I'm not saying it can't be done. I'm only giving our experience at attempting it. What we've found to be the main factor holding back total industry acceptance of remote sensing is how long it takes to extract useable data from point clouds. This is what I'm most interested in when it comes to solving this problem. A new fancy scanner is much less exciting to me than finally cracking the code to reliable automatic feature extraction.
As far as innovation goes, tx-surveyor has it right. Data collection is a breeze now. The future expertise will be developed in making the data that was collected useful. All of the value in a scanner resides in the individual capable of extracting accurate useful data.
The future expertise will be developed in making the data that was collected useful.
Amen.
Too bad so many firms are locked into "one-and-done" mode.
Or turning their noses up at anything other than boundary work because it's "not real surveying".
If you want tree identification. Look at hyperspectral. This allows you to categorize different signatures. Then extraction is easier to get it down to usable data that we can use on the surveying side. You can train your data extraction as well have a few random trees identified and located the old fashioned way aka survey it in. Then play with your bands to get the closest underlying signature of that species. Then you can run the overall extraction. To get the best bang for your buck. Okra and whacky tobacco have a very very close and similar signature. Had the boys checking out my garden once. Had to walk over and tell them they didn’t need to watch the house all night just come on up and see for themselves. He asked me how I knew. I said well I did this for a living in military and agency globally. lol. They did come look. Man that crimson okra was awesome that year.
Using our VLX 2 for interior building scanning is a game changer. If it requires a bit more accuracy, we will pull out the RTC360. The time difference is a lot though.
