View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0022999 | Open CASCADE | OCCT:Modeling Algorithms | public | 2012-02-17 14:58 | 2020-11-12 14:37 |
Reporter | Assigned To | bugmaster | |||
Priority | high | Severity | feature | ||
Status | closed | Resolution | no change required | ||
Product Version | 6.5.4 | ||||
Summary | 0022999: Implement method Normal for surface | ||||
Description | It is proposed to implement method Normal in the surface classes. Advantages are: performance for canonical geometry and more robust deifnition for some B-Spline cases | ||||
Tags | No tags attached. | ||||
Test case number | |||||
|
The normal is very useful especially for meshing algorithm. It's very hard to compute the normal at singularity points, for example the poles of sphere and cone. Looking forward to this feature. |
|
Dear Vico, Thank you for your remark. It is clear what to return in the case of pole of a sphere. But the question is what to return at pole of a cone? Possible variants are: - Normal to the cone surface corresponding to the angle of parameter U. - Direction of the cone axis (as average among all directions around the apex) - Null vector (to denote an ambiguity). The same question for a b-spline surface at points of various kinds of singularity: - Point at 1st derivative break. - Point where one of D1U or D1V is null (degenerated isoline). - Point where D1U and D1V are parallel (there are also such cases). |
|
Dear msv, It's really difficult to decide which variants to use. For me i'd like to return the average direation around the sigularity. It should not return null in any condition. |
|
If you use normal for creation of shading view of the model then using of average normal will give blurred (fuzzy) apex of the cone. |
|
Dear msv, You're right. So i think it should be better to return the normal corresponding to the angle of parameter U for cone surface. It's more difficult for b-spline surface. |
|
In order to compute a surface normal it is recommended to use the class GeomLProp_SLProps. If we add the method for computing normal in the Geom_Surface class we will need to do the same calculations. The questions what to do in case of singularities remain open, and I think they might be solved in different ways for different applications. So, I propose to close this bug as cannot be solved. |
Date Modified | Username | Field | Change |
---|---|---|---|
2012-02-17 14:58 |
|
New Issue | |
2012-02-17 14:58 |
|
Assigned To | => jgv |
2017-07-10 11:06 | kgv | Assigned To | jgv => msv |
2017-07-13 04:16 | Vico Liang | Note Added: 0068234 | |
2017-07-13 09:54 |
|
Note Added: 0068237 | |
2017-07-13 17:02 |
|
Status | new => assigned |
2017-07-13 17:35 | Vico Liang | Note Added: 0068273 | |
2017-07-13 18:36 |
|
Note Added: 0068276 | |
2017-07-14 06:06 | Vico Liang | Note Added: 0068298 | |
2020-11-12 13:02 |
|
Note Added: 0096700 | |
2020-11-12 13:02 |
|
Resolution | open => no change required |
2020-11-12 13:03 |
|
Assigned To | msv => bugmaster |
2020-11-12 13:03 |
|
Status | assigned => feedback |
2020-11-12 14:37 | bugmaster | Status | feedback => closed |