The V are the normalized V-residuals according to the formula:
q
TS (16)
x and s are computed with formulas (12) and the approximation values of
the unknowns which are denoted by 9 It 1s obvious that in any one specific
set of observational equations may appear either certain coordinate corrections
Ax or the corresponding residual errors V depending on the character of
the object point under consideration. With the introduction of the V-residuals,
according to formula (16), each ray has been assigned a specific group of such
residuals. Such an approach is desirable from the numerical standpoint, be-
cause any correlation is avoided between the coefficients of the matrix
associated with the residuals of the various rays, intersecting at a specific
point. From the theoretical standpoint such a solution seems to be advantageous
because the individual bundles of rays will conform to the pattern of the con-
trol data without undue restraint. In paragraph (D) of this chapter the possi-
bility will be discussed of arranging the least squares adjustment in such a
way that for each given control coordinate only one specific residual V is
obtained, independent of the number of rays intersecting at the point under
consideration.
The relation between the approximation values and the final values of the
unknowns is given by:
OQ +A% C me LAG
w = mn) + A Ww X = x? + AX
d p p
O O
K =K +AK = + A
Yo“ 7p Yo
O O +17)
XzX tX x = X + A&X.
o © o j 4 J
O 2 vO
Yo" Yg * ^ Y, Y Y: TAX
Oo o
= = + A
Zo 2, + A 2, 2, Zs Z,