注意R25 (也可能 R24)以上版本CATIA,使用 SPAWorkbench / Measurable将返回绝对坐标. 因此在R25 (也可能 R24)以上版本CATIA,以下矩阵运算可以省去.
There are two functions and a sub. The two functions are just mathematical stuff, 3x3 matrix inverse and determinant calculation, you can replace them for your preferred ones if you want. Sub Coord_Transform will do the coordinates transformation. The arguments of the sub are:
aRel() As Double , Matrix with a set of coordinates relative to the CATPart's axis system. These are, typically, the ones you get with the GetCoordinates method on a Point. Coordinates are as usual; aRel(0) for X, aRel(1) for Y and aRel(2) for Z. Maybe you need to convert types from variant to double to avoid type mismatch.
aAbs() As Double, Matrix with coordinates with respect to the Axis System of the top-level CATProduct. This is the matrix where you will get the results of the coordinates transformation. Initially blank, pass it to the sub and it will return it to you properly filled; again aAbs(0) for X, aAbs(1) for Y and aAbs(2) for Z
oProduct As Product, This is the CATPart's Product. For example, if you are measuring the coordinates of a "Point.1" in a CATPart named "Part.1", which is the first instance in a CATproduct named "RootProduct", then oProduct=RootProduct.Products.Item(1), and the CATPart itself can be accessed via the reference product, like oProduct.ReferenceProduct.Parent. You need to know the Product object which references to the CATPart in order to get its coordinates with the Position property.
bRecursively As Boolean, True if you want to calculate the coordinates with respect to the top-level Product, no matter what level of depth your CATPart is. False to calculate only with respect to the CATPart's father Product. I guess what you want to try here is "True", with some levels of depth...
Just paste the code in yours and call Coord_Transform with the appropiate arguments. If you need it, a similar routine also calculates coordinates taking into account different Axis Systems in a CATPart, using the same matrix transformations...
Hope it helps!
'---CODE START------------------------------------------------------' Function Det3x3(dX11 As Double, dX12 As Double, dX13 As Double, _ dX21 As Double, dX22 As Double, dX23 As Double, _ dX31 As Double, dX32 As Double, dX33 As Double) As Double '***********************************************' '*' '* 3x3 matrix determinant calculation (direct)' '*' '***********************************************' Det3x3 = dX11 * dX22 * dX33 + dX12 * dX23 * dX31 + dX21 * dX32 * dX13 - _ dX13 * dX22 * dX31 - dX12 * dX21 * dX33 - dX23 * dX32 * dX11 End Function Function Inv3x3(dX11 As Double, dX12 As Double, dX13 As Double, _ dX21 As Double, dX22 As Double, dX23 As Double, _ dX31 As Double, dX32 As Double, dX33 As Double, aInv() As Double) As Boolean '*********************************************** '* '* 3x3 matrix inverse calculation (direct) '* '*********************************************** Dim dDet As Double ReDim aInv(8) Inv3x3 = False dDet = Det3x3(dX11, dX12, dX13, dX21, dX22, dX23, dX31, dX32, dX33) If dDet = 0 Then Exit Function aInv(0) = (dX22 * dX33 - dX23 * dX32) / Abs(dDet) aInv(1) = (dX13 * dX32 - dX12 * dX33) / Abs(dDet) aInv(2) = (dX12 * dX23 - dX13 * dX22) / Abs(dDet) aInv(3) = (dX23 * dX31 - dX21 * dX33) / Abs(dDet) aInv(4) = (dX11 * dX33 - dX13 * dX31) / Abs(dDet) aInv(5) = (dX13 * dX21 - dX11 * dX23) / Abs(dDet) aInv(6) = (dX21 * dX32 - dX22 * dX31) / Abs(dDet) aInv(7) = (dX12 * dX31 - dX11 * dX32) / Abs(dDet) aInv(8) = (dX11 * dX22 - dX12 * dX21) / Abs(dDet) Inv3x3 = True End Function Sub Coord_Transform(aRel() As Double, aAbs() As Double, oProduct As Product, bRecursively As Boolean) Dim vProduct As Object, vCoord(11) Dim oFatherProduct As Product Dim aInv() As Double 'Exit condition, empty object If oProduct Is Nothing Then Exit Sub 'Redim absolute coords matrix On Error Resume Next ReDim aAbs(2) On Error GoTo 0 'Calculate product coordinates Set vProduct = oProduct vProduct.Position.GetComponents vCoord 'Calculate inverse matrix If Inv3x3(CDbl(vCoord(0)), CDbl(vCoord(1)), CDbl(vCoord(2)), _ CDbl(vCoord(3)), CDbl(vCoord(4)), CDbl(vCoord(5)), _ CDbl(vCoord(6)), CDbl(vCoord(7)), CDbl(vCoord(8)), aInv) Then Else MsgBox "Error, degenerate transformation", vbOKOnly Exit Sub End If 'Calculate transformation aAbs(0) = vCoord(9) + aInv(0) * aRel(0) + aInv(1) * aRel(1) + aInv(2) * aRel(2) aAbs(1) = vCoord(10) + aInv(3) * aRel(0) + aInv(4) * aRel(1) + aInv(5) * aRel(2) aAbs(2) = vCoord(11) + aInv(6) * aRel(0) + aInv(7) * aRel(1) + aInv(8) * aRel(2) 'If recursive option sepecified, search for parents and applies the transformation again If bRecursively Then 'Try to assign parent Set oFatherProduct = Nothing On Error Resume Next Set oFatherProduct = oProduct.Parent.Parent On Error GoTo 0 'If OK, recalculate coords If oFatherProduct Is Nothing Then Else aRel(0) = aAbs(0) aRel(1) = aAbs(1) aRel(2) = aAbs(2) Coord_Transform aRel, aAbs, oFatherProduct, True End If End If End Sub
As of R25 (or maybe R24) the SPAWorkbench / Measurable now always returns global point coordinates. So all the heavy matrix math is no longer required.