Contents   Index

Adjustment Report

At the end of the adjustment, a report on the adjustment is written to an output file with a .LST extension.
Cadastral Parcel Adjustment Job Name: T1
Zone Number is 561
Geodetic Corrections are applied
Direction (C-O) Limit is 2 13 20
Distance (C-O) Limit is 10.0

Transformation Results
Point X Y dx dy NAME
39 358850.000 1406950.000 -6.723 3.017 POR 46
7 360800.000 1406100.000 13.112 -10.819 POR 60
56 360800.000 1404500.000 -6.553 12.306 POR 43
38 358475.000 1405000.000 0.164 -4.504 POR 2

The dx,dy columns show how the new coordinates fit with the existing network. Any control points which have large residuals will be flagged with an asterisk and these values will not be used in the initial transformation but they will be used in the subsiquent adjustment. If there are any doubts with regard to any control points, terminate the adjustment at this stage. If only two control points are used, always check that both the scale factor and azimuth swing for the transformation are reasonable before proceeding.

Check Against Co-ordinates
Several checks are carried out on the way that the data fits in the adjustment model and warnings are printed for those items which are outside the angle and distance tolerance. These messages are written to the listing file. Observation equations are built from data in the parcels file, and as this proceeds, three types of warning messages may be given out.
1. C-O for bearings
2. C-O for distances
3. Line points off line

Transformation parameters are computed for each parcel to rotate, and scale it to the points data. After this overall rotation and scaling, the bearings of each line are compared to the bearings derived from the initial co-ordinates, and if the difference exceeds the limit, an error message is printed. Similarly if a scaled distance is different from the distance computed from co-ordinate data, a message is also given. For line points, a check is made to see if the sum of the two parts equals the total line length within the distance tolerance, and also if the point lies close to the line within the distance tolerance. If the point is out of tolerance, then the bearing and distance of the line and its two parts are displayed.

Warning Messages
The printout below shows examples of the three types of warning messages. The first 3 warnings are for direction observations. The next 3 lines are for a line point showing the total line and the two segments to the line points. The last 2 warnings are for distance equations and show the difference between the computed length of the line and the measured length of the line. It is up to the operator to decide whether to proceed at this stage, or to amend the data, and then proceed

Warning - Line 15 1 Direction C-O= -3 27 47
Warning - Line 1 22 Direction C-O= -3 53 53
Warning - Line 1 47 Direction C-O= -3 14 39
Warning - Check line point 36 on Line 37 35 40 21 55 903.05
Line 37 36 11 48 01 801.47
Line 36 35 102 54 25 431.88
Warning - Line 6 7 Distance C-O= 10.5
Warning - Line 7 12 Distance C-O= -12.0

If some large residuals are formed because points have been incorrectly joined, then the adjustment will tend to isolate the bad area and highlight the problem. If the cause is not obvious, it is reasonable to proceed, and then plot the adjusted network and inspect it together with the printed output for that section. All observation equations are used regardless of the size of their "C-O" values.

Adjustment
At the end of the equation forming phase, statistics for the job are displayed as shown below. The adjustment will proceed, the percentage complete is displayed on the screen as the equations are solved.

Number of Parcels:182
Number of Corners:492
Number of fixed points:4
Number of Directions:1193
Number of Distances:1011
Number of Observation Equations:2204
Matrix Size:96
Band Width:607
Number of terms in Normal Equations: 519536

Adjustment RESULTS
At the end of the adjustment, a report on the adjustment is written to an output file with a .LST extension.
First is a list of all the coordinates together with the corrections applied by the adjustment

A.C.S. COMPUTER GROUP Adjustment of Cadastral Parcels PAGE: 1
Geocadastre Ver 5.11 Sample_Job 19-Mar-2007
START TIME: 07:56:56
FINISH TIME: 08:34:11

AUSTRALIAN MAP GRID ZONE 561 - METRES
No. E (Final) N Corrections
1 386553.411 7379435.406 0.000 -.001
2 386574.128 7379440.500 0.000 -.001
3 386584.463 7379398.473 0.000 -.001
4 386557.826 7379391.922 0.000 -.001
5 386600.765 7379447.050 0.000 -.001
6 386611.099 7379405.023 0.000 -.002

Then there is an analysis of each parcel as shown below:

Plan OP1153 Parcel ID:366 Scale: 1.0000 Rotation 0 00 05
From Bearing Distance To DE DN Dist-Err
804 76 11 00 13.972 801 0.000 0.000 0.002
801 166 11 00 50.292 802 0.000 0.000 0.001
802 256 11 00 29.096 805 0.000 0.000 0.000
805 353 15 00 41.788 806 0.001
806 34 43 12 13.322 804 0.000
805 173 15 00 6.145 807 0.000 0.000 0.001

Comparisons made with each line in each parcel after a common scale factor and azimuth swing has been applied for all lines in the parcel. The differences between each line computed from its adjusted coordinates, and that from the rotated and scaled parcel dimension is then displayed. The “Dist-Err” is the difference between the measured distance and the length computed from the adjusted coordinates after the projection line scale factor has been applied. Points which have been excluded from the adjustment are then computed, and a linear adjustment is used for points on traverse loops which have been replaced with a single vector in the adjustment. In the above example, there are no other lines connecting into point 806 so the lines 805-806 and 806-804 were replaced in the adjustment by a single line and therefore there were no “DE DN” comparisons for those lines in the report. As the parcels are printed the mean standard error for line lengths is computed and those lines which have residuals exceeding two standard deviations are printed out together with the mean standard error for the set. This data gives an indication as to the accuracy of the data as well as flagging those lines which are suspect. At the end of the listing those lines exceeding two standard deviations from the standard error are listed together with the parcels they belong to. This provides a method of rapidly detecting the main errors in the job.

** Points with a error exceeding two sigma **
PLAN Parcel Parcel Misclose Point dX dY East North
752459 22 0.000 0.000 270.213 0.167
752317 23 0.016 0.002 260.105 0.111
752317 24 0.025 0.013 300.005 0.213

Sigma for C-O = 0.001

The standard deviation of all errors is also printed out as an indicator as to the precision of the adjustment.

TROUBLE SHOOTING
If a parcel is connected to the edge of an area by a minimal connection (such as one corner and one line point) then it can cause the adjustment to fail, and the point being processed at the time of failure will be printed. There are several ways to overcome this type of problem and they include:
1.Remove the offending parcel
2.Provide more connections to the offending parcel
3.Place a control point closer to the parcel

Parcels may sometimes appear to be joined when viewed on the screen, but in fact may not be joined. This type of error can be detected by using the “Mean Two Points” tool see Mean Two Points

Adjustment Method
When building the adjustment, the program eliminates lines which do not contribute to the strength of the network. Connections which have not been used are excluded, and multiple boundary lines between node points are replaced by a computed bearing and distance from the lines. After the adjustment, these points are interpolated from the surrounding adjusted points
Easement lines (coded 900 - 994) are excluded.
Lots tagged as compiled plans are set to the lowest weighting.
Lines tagged with code 995 are given a weighting one above that of their parcel These lines are generally used for reliable connections to control points from the cadastre.
Initial values for the coordinates are the current values transformed into the control point system.
A distance equation and a direction equation is written for each line, and two direction equations are written for each line point. The observation equations have the same structure as those used in the adjustment of a geodetic network, except that all lines in a parcel are treated as a single direction set, and a single orientation coefficient is provided per parcel.
Adjustment is by variation of coordinates using a banded matrix. Normal equations are formed from the observation equations built earlier, and these are stored on a temporary file on disk before being solved by a Cholesky solution.