Bad GPS location filter
Gathering feedback
Hello,
Recently on this topic
http://help.locusmap.eu/topic/use-srtm-data-from-hgt-files-for-height
we also talk about how bad recorded points affected the statistics of speed and altitude.
Now I ask if there is the possibility to introduce a filter for bad points.
The filter can intercept those points for example too far from the previous one recorded: if you are not Usain Bolt you cannot run 100 Mt in less than 10 sec, and also you cannot climb faster than Denis Urubko.
Sometimes my gps makes me think I am a superman but I have to fight with my sad reality of a limited human being and bad recorded points.
I think the filter should be based on activities due to different speed involved.
I like this idea 
This is my absolute top priority issue with Locus. Usually the GPS of my phone is reasonable accurate, but sometimes close to rock walls or similar some points get recorded which are way off. I'd love to have a simple plausibility filter which discards points which are off hundreds of meter in altitude or lateral position and jump back after several seconds to the correct position.
This is my absolute top priority issue with Locus. Usually the GPS of my phone is reasonable accurate, but sometimes close to rock walls or similar some points get recorded which are way off. I'd love to have a simple plausibility filter which discards points which are off hundreds of meter in altitude or lateral position and jump back after several seconds to the correct position.
There has been various attempts to address this issue, none 100% reliable of course. including NMEA VDOP field threshold.
There has been various attempts to address this issue, none 100% reliable of course. including NMEA VDOP field threshold.
Although it is probably not so simple to design a filter that filters the live position stream it should be much easier to clean a recorded track. After all one only has to look for large (relative to the median distance between subsequent track points) jumps in coordinates that jump back close to the position before the jump a short while after. I could even think of having it as part of the track editor where one gets a preview of the to be modfied parts of the track.
Although it is probably not so simple to design a filter that filters the live position stream it should be much easier to clean a recorded track. After all one only has to look for large (relative to the median distance between subsequent track points) jumps in coordinates that jump back close to the position before the jump a short while after. I could even think of having it as part of the track editor where one gets a preview of the to be modfied parts of the track.
Replies have been locked on this page!