Kris translate this and try to understand what want to say to you, to you career in Photogrametry..
Kris, the images are RENDERED out from MAX as i get. Therefore the background is empty space... so you test data that are generated by software, no lens distortion present ( can be generated artifitialy ) so we can say is a "non realistic"
Try understand that EVERY sw count with some lens distortion, if you remove it, the sw get some level of problems - every lens create it, therefore it uses its own specialy "constructed" algorithm for removing it. SW like lightroom or pshop use just very simple models of removing distortion - read about it that you understand it better..
Is writen that scenes that are flat - that are like this scene, is problematic in EVERY photogrametry solution........
Tested to align, it get just about 2800 points on 16 photos, that is realy small amount for proper work.. ideal is around 1000 per image ( whitepaper about it ) in simple scenes. In normal realworld scene you get over 25.000+ - much precise data collected for pscan eingine. Some of us use a special workflow, that object like this are extra painted, some degree of adding data that sw can better align it....
As were mentioned in another thread, to work eficiently with photogrametry is need to know some "magic" or better to say is a ART of generating proper and good models, that you get if work a longer period with it, that you understand of how it work, what bring better results, what worse.....
Do you have ever tested Photomodeler Scanner that was defacto standard in photogrametry over almost a decade ? Ther you must shoot in very precise manner, under very tight conditions, if not then you get no data..... a lot of us know that sw package and problems to create results that could be used..
Please for your personal interest read a few whitepapers about Photogrametry that you understand better and deeper how it work. Therefore a lot of people who started with it are confused that other getting excelent results and not getting the ideal workflow to its own data and blaming that is just a propaganda...... most of us have developed skills in particular areas, but it cost us a LOT of homework a lot of hours of testing and therefore we want not to give it freely, that we give it our competition..
Not everything can be reconstructed idealy, some areas are hard or better not possible to create, mostly scenes with a lot of flat ares - no enought points to align and other things....
So try to rethink your workprocess, you do something bad that we clearly see, but you cannot understand it, becasue lack of knowlege.... Don?t get it that im want war with you, im just try to give you hints how to work efficiently with this software........ but we wil not give you step by step solution, we earn money for our work so learning someone our secret is giving them a better start agains us...... In my country few have found pscan afther im put results online, they want take my clients, but they cannot bring results like me, becasue they think is a soo simple and not understand the background ( and they are in bussines longer than 10 years ), my own hours of testing comparing writing down and implementing in my workflow....
My personal recomendation is to READ few books about photogrametry then with your knowlege of other things can create a lot of nice outputs, but without the background is usseles to explain a lot of things.....
Will contact you over Skype later.....
Topic: AGISOFT Laboratory Test Procedure. (Read 12728 times)
