| Using "tie points" in rasters [message #93199] |
Wed, 11 May 2016 14:35  |
GB
Messages: 2
Registered: May 2016
|
Junior Member |
|
|
Hello,
I am working with a digital elevation map raster and ran into a problem that I am unsure how to approach.
Once I get the spatial reference object of my DEM raster, I then get the tie_point_pixel, the tie_point_map, and pixel_size from the spatialref. I then seek to transform the map coordinates of a set of LiDAR points into the pixel indices of my raster. The thing I am confused about is how rounding plays into this problem.
For instance, lets say that I use the following function to convert the x-coordinate of my lidar point into pixel coordinates:
(lidar_point_x-tie_point_map[0])/pixel_size[0]
But, I get a decimal place in the outcome. Lets say that the decimal point is 0.64. Would this decimal point translate into the point belonging to index 0 or to index 1?
I suppose I am confused about if the tie_point_map is centered on the tie_point_pixel such that it extends a range of -0.5*pixel_size to 0.5*pixel_size. If anyone could elaborate on this I would be very thankful.
|
|
|
|
| Re: Using "tie points" in rasters [message #93202 is a reply to message #93199] |
Wed, 11 May 2016 20:21  |
jiashenyue
Messages: 10
Registered: January 2013
|
Junior Member |
|
|
On Wednesday, May 11, 2016 at 2:35:04 PM UTC-7, GB wrote:
> Hello,
>
> I am working with a digital elevation map raster and ran into a problem that I am unsure how to approach.
>
> Once I get the spatial reference object of my DEM raster, I then get the tie_point_pixel, the tie_point_map, and pixel_size from the spatialref. I then seek to transform the map coordinates of a set of LiDAR points into the pixel indices of my raster. The thing I am confused about is how rounding plays into this problem.
>
> For instance, lets say that I use the following function to convert the x-coordinate of my lidar point into pixel coordinates:
> (lidar_point_x-tie_point_map[0])/pixel_size[0]
> But, I get a decimal place in the outcome. Lets say that the decimal point is 0.64. Would this decimal point translate into the point belonging to index 0 or to index 1?
>
> I suppose I am confused about if the tie_point_map is centered on the tie_point_pixel such that it extends a range of -0.5*pixel_size to 0.5*pixel_size. If anyone could elaborate on this I would be very thankful.
Usually I just round the decimal points to the closest pixel. If I got a index of 1.64, I go for 2. I would say this is quite fair, since in this case, the pixel you are trying to locate is closer to the next pixel.
Shenyue
|
|
|
|
| Re: Using "tie points" in rasters [message #93205 is a reply to message #93202] |
Thu, 12 May 2016 05:52 |
GB
Messages: 2
Registered: May 2016
|
Junior Member |
|
|
On Wednesday, May 11, 2016 at 11:21:31 PM UTC-4, SHENYUE JIA wrote:
> On Wednesday, May 11, 2016 at 2:35:04 PM UTC-7, GB wrote:
>> Hello,
>>
>> I am working with a digital elevation map raster and ran into a problem that I am unsure how to approach.
>>
>> Once I get the spatial reference object of my DEM raster, I then get the tie_point_pixel, the tie_point_map, and pixel_size from the spatialref. I then seek to transform the map coordinates of a set of LiDAR points into the pixel indices of my raster. The thing I am confused about is how rounding plays into this problem.
>>
>> For instance, lets say that I use the following function to convert the x-coordinate of my lidar point into pixel coordinates:
>> (lidar_point_x-tie_point_map[0])/pixel_size[0]
>> But, I get a decimal place in the outcome. Lets say that the decimal point is 0.64. Would this decimal point translate into the point belonging to index 0 or to index 1?
>>
>> I suppose I am confused about if the tie_point_map is centered on the tie_point_pixel such that it extends a range of -0.5*pixel_size to 0.5*pixel_size. If anyone could elaborate on this I would be very thankful.
>
> Usually I just round the decimal points to the closest pixel. If I got a index of 1.64, I go for 2. I would say this is quite fair, since in this case, the pixel you are trying to locate is closer to the next pixel.
>
> Shenyue
That is what I was thinking. I was just confused about whether or not the tie_point_map was centered on the pixel or denoted the upper left-hand corner of the pixel.
Thanks for the help.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
