TxDOT Survey Manual: Photogrammetry

Section 4: Photogrammetry

Control Target Documentation

All control points will be documented with:

a graphical representation of the location of markers
a textual documentation in the form of a spreadsheet.

The above information shall be provided to the TxDOT District Survey Coordinator. However, the TxDOT District Survey Coordinator may require additional documentation depending on the requirements of the project.

Graphical documentation presented as a sketch or CAD drawing shall include the following information.

Graphical information:
spatial location of control point with sufficient detail to allow recovery of the point in the field
map scale and north arrow
stationing annotated project centerline or roadway alignment (if available) as supplied by TxDOT.
Textual information:
date of installation
highway name and project limits
general description of the location of point referenced to highway intersections, city limits, etc.
TxDOT work order number and control section job (CSJ) number (if applicable)
identification of reference coordinate system and elevation datum
combined surface adjustment (CAF) or TxDOT Surface Adjustment Factor (SAF)
point coordinates identified as being either grid or surface coordinates.

Textual documentation shall be submitted using the Microsoft Excel spreadsheet titled Ground Control Submission Form (2154). See Figure 3-1 below for an example of the form.

Note that the standard deviations computed from the network adjustment shall be included on this form.

Figure 3-1. Ground Control Submission Form (2154).

Second page of Form 2154. The above form may be access through the internal TxDOT Forms System. Contractors should request this form from the TxDOT District Survey Coordinator.

Anchor: #i1110650

Control Targets - General

The secondary control network described in the previous section is marked with cross-shaped targets during the aerial photography flight. The targets are visible on the developed aerial film and are used to relate the aerial photography to the ground. Targeting is an essential part of photogrammetric mapping. Studies conducted by the Federal Highway Administration (FHWA) determined that pre-marked targets improve the efficiency of photogrammetric processing and increase overall accuracy of point measurements.

A right-of-entry agreement with a landowner must be obtained prior to entering private property to set photogrammetric control. The agreement shall include detail on the length of time that the control target will be on the ground and those responsible for removal of the material. The agreement may include additional detail such as specific times for access to the property or conditions on notification before entering the property.

Targets may be placed on hard surfaces such as pavement or concrete or on soft surfaces such as soil, gravel, or grass. In general, hard surfaces are preferred because point measurements both for the surveyor and for the photogrammetrist, are more accurate on a hard surface. In addition, a hard surface allows targets to be painted which increases the durability and longevity of the mark.

The cross-shaped target is centered on the PK nail, iron rod, or other surveying marker that defines the control point. On hard surfaces the targets shall be painted using flat finish paint. The target color shall be either white (on darker background surfaces) or black (on lighter background surfaces). Contrast between the target and the background material is important for point measurement in the photogrammetric process. It is permissible to use both paint colors, one as a background color, and the other as the target color to enhance the contrast further. However, in general, white targets are preferred.

On soft surfaces the target is constructed of cardboard, sheetrock, plastic film, Tyvek, or other similar materials durable enough to remain in place until the flight mission is complete and the film has been inspected. Once the flight mission has been approved, the target materials shall be picked up.

Placement of control targets shall meet the following criteria:

The target should be clear of any obstruction that may obscure the target on the aerial photography. When standing on the target, there should be a clear view of the sky from 45^o above the horizon to zenith in every direction.
The target shall be placed as flush to the ground as possible. Any vegetation that may grow beneath the target should be cleared prior to placing the target material. Vegetation growth can cause a target to bow, affecting the accuracy of the photogrammetric measurement.
Targets should not be placed in a shadow or in a shady area. It is advisable to visit target locations at the approximate time of the flight mission to determine the location of shadows.
Placement of targets beneath overhead wires should be avoided. Overhead wires make measurement of the targets in the photogrammetric process problematic.
Targets should be placed on as level an area as possible. If a target must be placed on a slope, the target should be oriented so that two of the legs that form a straight line traverse the slope at approximately the same elevation.
The horizontal coordinate provided for the target shall be at the center of the target at the location of the control monument. The vertical elevation of the target shall be at the center of the target at the level of the target material. This is important because the photogrammetric measurement of the point is made at the elevation of the target material, which may or may not be the elevation of the survey marker.

Control Targets - Design

The following figure (Figure 3-2) shows the design of a typical control target. Note that the size of the target is dependent on the scale of the aerial photography for which it will be used. Target design and dimensions.

Figure 3-2. Target Design and Dimensions.

For a typical 1'' = 250' photo scale the dimensions of the target are: Thickness (T) = 6 inches and Length (L) = 84 inches. For projects with different photo scales, consult with the photogrammetrist for the correct target dimensions.

The illustration below shows the design of a box type target used to designate new flight tangents.

Figure 3-3. Box Target Design.

A flight tangent is a single, straight flight line maintained by the aircraft. Anytime the aircraft has to make a turn to maintain photo coverage over a roadway, that turn and new flight direction constitutes a new flight tangent. Generally, a point of intersection (PI) in a roadway of 15 degrees or more will require a new flight tangent.

Control Targets—Placement

Placement of the secondary control network survey markers and targets shall adhere to the following requirements:

A band of control is placed 300 to 500 feet before the beginning and ending control bands for the project. A band of control is defined as being the centerline survey marker and its associated wing markers.
Control bands are placed at 1500' maximum and 1000' minimum intervals along the flight line. The maximum distance between bands can be increased to 1800 feet but not for two or more consecutive bands.
Wing markers are set at the width of the mapping area with a maximum distance of 600 feet from the centerline. The minimum distance is typically 200 feet from the centerline marker. However, in situations where access to private property is denied or physical features prevent placement at the minimal of 200 feet, target can be placed at the edge of the right-of-way.
The first and last centerline survey point of each separate flight line shall be constructed as a “box” marker. Note that if the roadway being covered by the aerial photo mission has a PI greater than 15 degrees, a new separate flight line will be required to maintain photo coverage. In this case the new flight line will require box panels at the beginning and ending of the flight line.
All flight lines require a minimum three (3) horizontal markers.

The figure below illustrates the control marker and target layout for a typical aerial mapping project. Note the PI that occurs towards the center of the figure and the corresponding control point locations.

Figure 3-4. Example of Control Target Layout and Naming.

Photogrammetric Mapping

I. Aerial Photography - General

Film titling should consistently be on the left side of the image frame regardless of the flight direction. The title will include the aerial photo date, the project SRN number (or the CSJ number or project number as directed by the TxDOT District Survey Coordinator), the highway or project name, and the photo strip and frame number. An example of film titling is provided below:

Figure 3-5. Example of Film Titling.

All aerial film is the property of TxDOT. Film shall be delivered to the TxDOT District Survey Coordinator as a project deliverable. The film will be archived at the TxDOT Technology Services Division (TSD).

II. Aerial Mapping - General

Equipment used in the aerial mapping process is required to produce maps conforming to American Society of Photogrammetry and Remote Sensing (ASPRS) Class 1 Accuracy Standard. This requirement affects all equipment and workflows used in the mapping process including: aircraft operation, the aerial mapping camera, the film scanner, and the photogrammetric workstation. This requirement also affects operator proficiency for all systems and workflows involved in map production.

TxDOT requires that maps produced photogrammetrically be delivered in MicroStation ® V8 file format. The following level structure is required for the maps.

The following items are the standard deliverables for an aerial mapping project:

Anchor: #i1142604Table 3.5 Photogrammetric Level Structure
Photogrammetry Feature	DTM	Microstation V8 Name	Level
Control
Horizontal Control, Principal Point	no	p_control ground ctrl	401
Road
Paved Road, Curb	yes	p_road paved, curb	402
Dirt Road	yes	p_road dirt	403
Guard Rails	no	p_road guard rail	404
Guard Fences	no	p_road guard fence	405
Guard Posts	no	p_road guard post	406
Concrete Barrier	no	p_road conc barrier	407
Paint Stripe Solid and Dashed	yes	p_road paint stripe	408
Bridge End	yes	p_road bridge end	409
Cattle Guard	no	p_road cattle guard	410
Overhead Sign	no	p_road overhead sign	411
General Road Feature	no	p_road general feature	412
Railroad
Railroad Track RR Controls	no	p_railroad rr control	413
Drainage
Concrete Dam	yes	p_drainage conc dam	414
Concrete Drain	yes	p_drainage conc drain	415
Earthen Dam	yes	p_drainage earthen dam	416
Riprap	yes	p_drainage riprap	417
Culvert	yes	p_drainage culvert	418
Inlet	yes	p_drainage inlet	419
Water	yes	p_drainage water	420
Marsh	yes	p_drainage marsh	421
Structure
Building	no	p_structure building	422
Ruin	no	p_structure ruins	423
Sidewalk	no	p_structure sidewalk	424
Slab	no	p_structure slab	425
Porch, Deck	no	p_structure porch	426
Stairs, Steps	no	p_structure stairs	427
Fence, Gate, Post	no	p_structure fence	428
Retaining Wall	no	p_structure ret wall	429
Wall	no	p_structure wall	430
Cemetery	no	p_structure cemetery	431
Billboard	no	p_structure billboard	432
Sign, Sign Pole, Sign Post	no	p_structure sign	433
Antenna, Cellular Tower, Satellite Dish	no	p_structure antenna	434
Windmill	no	p_structure windmill	435
Flag Pole	no	p_structure flag pole	436
Pipes	no	p_structure pipe	437
Tank	no	p_structure tank	438
Area Under Construction	no	p_structure constr area	439
General, Ac Unit, Goal, Large, Small Circle	no	p_structure general	440
Unidentified Feature	no	p_structure unidentified	441
Utility
Fire Hydrant	no	p_utility fire hydrant	442
Manhole	no	p_utility manhole	443
Marker, Meter, Valve	no	p_utility marker	444
Transmission Tower, transmission Line	no	p_utility trans tower	445
Pipeline	no	p_utitlity pipeline	446
General, Pole, Pole LP, TFP, LP
Traffic Light, Gas Light	no	p_utility general pole	447
Vegetation
Woods	no	p_veg woods	448
Tree	no	p_veg tree	449
Tree Farm	no	p_veg tree farm	450
Tree Orchard	no	p_veg tree orchard	451
Palm	no	p_veg palm	452
Digital Terrain Model (DTM)
Breakline	yes	p_dtm breakline	453
General Breakline	yes	p_dtm general breakline	454
Retaining Wall Breakline	yes	p_dtm retaining wall	455
Sidewalk Breakline	yes	p_dtm sidewalk	456
Mass Points	yes	p_dtm mass points	457
Water Obscured	yes	p_dtm water obscured	458
Obscured Area	yes	p_dtm obscured area	459
Pit and Fill Area	yes	p_dtm pit or fill area	460
Stock Pile	yes	p_dtm stock pile	461

The following items are the standard deliverables for an aerial mapping project:

Aerial film negatives delivered on the file roll to the TxDOT District Survey Coordinator. The film will be archived at TSD.
Contact prints as required by the TxDOT District Survey Coordinator.
Control point documentation.
MicroStation^® planimetric file as a 2D design file.
MicroStation^® DTM data file as a 3D design file.
GEOPAK^® TIN file.

If a field check (see subheading, “Accuracy Specification” below) is performed and additional deliverable of the statistical computation sheets is required.

II. Map Accuracy Standard

The TxDOT map accuracy standard for maps created using aerial photography is Class 1 of the American Society for Photogrammetry and Remote Sensing (ASPRS) Specifications and Standards Committee Standards for Large-Scale Maps (ASPRS 1990).Map accuracy depends on all of the individual components that go into the mapping process: ground control accuracy, precision, and spatial geometry; aerial photo acquisition, quality, and processing; analytical triangulation equipment and procedures; and map compilation equipment, procedures, and accuracy. A mapping project can be designed with processes and procedures in place to meet a particular map accuracy specification. However, actual map accuracy can only be determined by an independent field check of the map using surveying equipment and procedures that result in higher accuracy coordinate determination than those of the map being checked.The TxDOT map accuracy standard allows production of photogrammetrically derived maps to adhere to the Class 1 ASPRS standard without the necessity of a field check provided the proper equipment and procedures are used. However, periodically TxDOT may request a field check on a map or may perform a field check using TxDOT equipment to ensure that the required accuracy standards are being met.

Accuracy Specification

The standard places limits on the root mean square error (RMSE) for individual position components - northing, easting, and elevation.The limiting horizontal RMSE for large scale (1:20,000 or larger) Class 1 maps is 0.01'' at map scale. Therefore a map compiled to a scale of 1'' = 50' (1:600) would have a limiting RMSE for the X and Y components of 0.5'.The following table gives the horizontal limiting RMSE for some common maps scales:

Anchor: #i1152712Table 3.6 Horizontal Limiting RMSE
Limiting RMSE Feet	Map Scale - Imperial	Map Scale Ratio	Practical Uses
0.05	1'' = 5'	1:60	Typical Ground Methods
0.1	1'' = 10'	1:120
0.2	1'' = 20'	1:240
0.3	1'' = 30'	1:360	Typical Photogrammetric Methods
0.4	1'' = 40'	1:480
0.5	1'' = 50'	1:600
1.0	1'' = 100'	1:1,200
2.0	1'' = 200'	1:2,400

The limiting vertical RMSE for large scale (1:20,000 or larger) Class 1 map elevation data is 1/3 of the indicated contour interval. Spot elevations are restricted to an RMSE of 1/6 of the contour interval. A map compiled for a one foot contour interval would therefore have a limiting RMSE for elevation data of 0.33'. The limiting RMSE for spot elevations would be 0.17'.

Limiting RMS Error

Limiting RMS error places a limit on the statistical value for a sampling of differences in X, Y, and Z coordinate values between features on the photogrammetrically compiled map and those identical features surveyed in the field using equipment with a higher level of positional accuracy.

Limiting RMS error should not be confused with the maximum allowable error for any individual coordinate component. It is possible for individual differences to exceed the limiting RMS error while the overall RMSE meets specification. Mathematically, the limiting RMSE is determined using the following equation:

where:D² = d₁² + d₂² + d₃² + ….. + dn²

d = discrepancy in the X, Y, or Z coordinate direction = x_map - x_check

n = total number of points checked on the map in the X, Y, or Z coordinate direction

Map Accuracy Testing

Tests for compliance of a map are optional but are recommended if the accuracy isn’t reported by the data provider or if the map accuracy is questionable. TxDOT may request a map accuracy test from a vendor for any photogrammetrically derived map. Alternatively, TxDOT staff may perform an accuracy check using state-owned resources. A written summary of the results of the field check is required if a vendor is asked to perform the map check.

The written summary will include the project SRN number, the roadway name being mapped, the county the project is in, and the date that the field check was performed. A brief summary of the field procedures and equipment used shall be included in the summary.

A table shall be included in the summary which shows the:

point number and description of the check point
X map coordinate, the X field surveyed coordinate
difference between the two coordinates
difference squared.

The same data should be included for the Y coordinate. A separate table with the same information should be prepared for the Z coordinate. At the bottom of the difference squared column, the average of the squares and the square root of the average shall be shown and labeled. The square root of the average of the squared differences is the limiting root mean square error (RMSE) value. This is the value that will be used to determine if the map meets the required accuracy specification.

Testing for horizontal accuracy compliance is done by comparing the planimetric (X and Y) map coordinates of well-defined points on the ground to the coordinates of the same points as determined by a horizontal check survey of a higher accuracy. The check survey shall be designed to achieve standard deviations equal to or less than one-third of the RMSE selected for the map.

Testing for vertical accuracy compliance shall be accomplished by comparing the elevations of well-defined points from the map to corresponding elevations determined by a survey of higher accuracy. For purposes of checking elevations, the map position of the ground point may be shifted in any direction by an amount equal to twice the limiting RMSE in position. The vertical check survey should be designed to produce RMSE in elevation differences at check point locations no larger than 1/10^th of the indicated contour interval.

A minimum of 20 check points shall be used for field checking in order to produce acceptable statistical results. Check points should be evenly distributed throughout the entire map area. Discrepancies between X, Y, or Z coordinates of the ground point, as determined from the map and by the check survey that exceed three times the limiting RMSE shall be interpreted as blunders and will be corrected before the map is considered to meet this standard. All points determined to be blunders either on the map or as part of the field check survey shall be documented in the map check summary.

The following figure shows typical statistical check sheets for horizontal and vertical accuracy of an aerial derived map.

Figure 3-6. Example of Horizontal and Vertical Map Check Statistical Computations.

Statement of Map Accuracy

For maps that are not field checked but have been compiled to meet TxDOT’s accuracy specification, the following statement shall be included on the delivered hardcopy and digital versions of the map:

“THIS MAP WAS COMPILED TO MEET THE ASPRS STANDARD FOR CLASS 1 MAP ACCURACY”

If the map was checked and found to conform to this spatial accuracy standard, both the statement above and the following statement shall be included on the delivered hardcopy and digital versions of the map, as well as in the field check summary:

“THIS MAP WAS CHECKED AND FOUND TO CONFORM TO THE ASPRS STANDARD FOR CLASS 1 MAP ACCURACY”

Step-by-Step Accuracy Testing

Select the check points. Select a minimum of 20 well-defined points in the map dataset. A well-defined point represents a feature for which the horizontal position is known to a high level of accuracy. For the purpose of accuracy testing, well-defined points must be clearly defined on the map product and easily recoverable on the ground. Check points shall be evenly distributed throughout the area covered by the map.
Field survey the check points. Using a survey method of higher accuracy (typical optical or GPS surveying equipment) occupy the selected check points and record the X, Y, and/or Z value for each point. Note that a single point can be used to check both planimetric and elevation accuracy. The survey method should be capable of measuring with a horizontal accuracy one-third the selected limiting RMSE value for the map. The survey should also be capable of a vertical measurement 1/10th of the indicated contour interval of the map.
Document the horizontal test. Produce a table listing: Point number, description, X coordinate determined from the source of higher accuracy, X coordinate determined from the map, the difference between the two coordinates, the squared difference, Y from the source, Y from the map, the difference, the squared difference. This table will be included in the field check summary documentation.
1. Sum the squared differences and compute the average. From the table column containing the X squared difference, sum the values and compute the average difference. Do the same for the column containing the Y squared difference values.
2. Calculate the root mean squared error. Use the root-mean-square error (RMSE) to determine the positional accuracy in both the X and Y coordinate direction. RMSE is the square root of the average of the set of squared differences between dataset coordinate values and coordinate values from an independent source of higher accuracy for the identical points. Take the square root of the average computed in the previous step for both the X and Y columns. This value is the limiting RMSE for the map. This value should not exceed the limiting RMSE specification for the map scale being checked. If the computed RMSE exceeds the limiting RMSE value for the map scale, the map fails its accuracy specification. The equation for the RMSE test is as follows:
where:

D² = d₁² + d₂² + d₃² + ….. + d_n²

d = discrepancy in the X or Y coordinate direction = X_map - X_check

n = total number of points checked on the map in the X or Y coordinate direction
Document the vertical test. Produce a table listing the: Point number, description, Z coordinate determined from the source of higher accuracy, Z coordinate determined from the map, the difference between the two coordinates, and the squared difference. This table will be included in the field check summary documentation.
1. Sum the squared differences and compute the average. From the table column containing the Z squared difference, sum the values and compute the average difference.
2. Calculate the root mean squared error. Use the root-mean-square error (RMSE) to determine the positional accuracy in Z coordinate direction. RMSE is the square root of the average of the set of squared differences between dataset coordinate values and coordinate values from an independent source of higher accuracy for the identical points. Take the square root of the average computed in the previous step for the Z column. This value is the limiting RMSE for the map. This value should not exceed the limiting RMSE specification for the map scale being checked. If the computed RMSE exceeds the limiting RMSE value for the map scale, the map fails its accuracy specification. The equation for the RMSE test follows:
where:

D² = d₁² + d₂² + d₃² + ….. + d_n²

d = discrepancy in the Z coordinate direction = Z_map - Z_check

n = total number of points checked on the map in the Z coordinate direction
Prepare the summary. The written summary will include the project SRN number, the name of the roadway being mapped, the county that the project is in, and the date that the field check was performed. A brief summary of the field procedures and equipment used should be included. The tables prepared from the data listed above shall be included in the summary. The table will include the point number and description of the check point, the X _map coordinate, the X field coordinate, the difference between the two points, and the difference squared. The same data should be included for the Y coordinate. The table prepared from the data listed above for the Z coordinate check shall also be included in the summary. At the bottom of the difference squared column, the average of the squares and the square root of the average should be shown and labeled. The square root of the average of the squared differences is the RMSE value.