Fish Track Detection Configuration Properties dialog box
The Fish Track Detection Configuration Properties dialog box is used to view and enter the settings that will be used for the fish track detection process. Echoview supports fish track detection configurations, where different detection settings can be specified for different data or analysis requirements.
A single target represents an echo from a fish at a particular ping time and range or depth and, where applicable, position within the beam. The detection properties control how Echoview predicts and matches the movement of single targets to identify fish tracks. For more information, see Fish tracking algorithms.
To display the Fish Track Detection Configuration Properties dialog box:
- Display the EV File Properties dialog box.
- Click the Region Detection page.
- Under Fish tracks, select a detection configuration, then click Properties...
- OR -
- Display the Detect Fish Tracks dialog box.
- Click Edit Region Detection Configurations...
- Under Fish tracks, select a detection configuration, then click Properties...
There are three sections on the Fish Track Detection page. Each section represents a stage in the fish track detection process. See Fish tracking algorithms for a more detailed description of the processes followed in Echoview.
Fish Track Detection
Algorithm
The Algorithm section is used to identify which single targets are candidates for adding to a fish track.
Data
The Data sub-section is used to choose one of two algorithms available for fish track detection.
|
Settings |
Description |
|
2D (range and time) |
Choose this option for data from a single beam or dual beam echosounder. You can also choose this option for split beam data when you want the detection algorithm to disregard the angle data. All targets are then treated as though they were at the center of the beam. |
|
4D (range, angles and time) |
Choose this option for split beam data when you want the detection algorithm to use angle data—the target’s position within the beam—as well as range and time. The quality of the angle data is important for 4D detection. Consider the effects of noise and reduced angle-data precision. |
Track detection
These settings are used to determine a predicted location for the next target in a track. This predicted location is compared with the observed location of each target to determine whether the target is a candidate for the track.
|
Settings |
Description |
|
Gain settings that define the algorithm's sensitivity to unpredicted changes in position. The basic tracking quantities are Major axis angle, Minor axis angle, and Range. Lower Alpha values are less sensitive to measurement noise; higher Alpha values are more responsive to target movement. |
|
Beta |
Gain settings that define the algorithm's sensitivity to unpredicted changes in velocity, including changes in the direction or speed of a target. The basic tracking quantities are Major axis angle, Minor axis angle, and Range. Lower Beta values are less sensitive to measurement noise; higher Beta values are more responsive to changes in target movement. Velocity is generally more sensitive to noise than position, so Beta is generally lower than Alpha. |
Tip: For a target that swims or travels in a highly predictable fashion low gain values should be set. Conversely, the more erratic the target's swimming behavior the larger the gain settings that are required.
See About beam geometry for a description of the axes.
Target gates
Target gates are used to determine whether the observed location of a target is close enough to the predicted location for the target to be considered as a candidate.
|
Settings |
Description |
|
Exclusion distance (m) |
Choose distance(s) around the predicted location within which a target must be located in order to be a candidate. The basic tracking quantities are Major axis angle, Minor axis angle and Range. Hint: The Exclusion distance should be of the same order of magnitude as the expected changes in position. |
|
Missed ping expansion (%) |
Increases the exclusion distances for each missed ping between the last target in the track and the current ping. The expansion is linear (or arithmetic) for successive missed pings. The basic tracking quantities are Major axis angle, Minor axis angle and Range. To disable the missed ping expansion, accept the defaults of zero. |
For 2D fish track detection, the difference in range between the predicted and observed location of the target must be within the Range target gate.
For 4D fish track detection, the settings are used to create a volume around the predicted location within which the observed location must fall. See About beam geometry for a description of the axes.
Notes:
- The target gate is calculated as: Exclusion distance + N × (Exclusion distance × Missed ping expansion/100), where N is the number of missed pings.
- Exclusion settings on the Analysis page of the Variable Properties dialog box also affect fish track detection.
Weights
Weights are used for the weighting and allocation of single target candidates to a track.
Once a single target is identified as a candidate for a track, a measure of its suitability as a candidate for the track is determined by considering the weights of particular "track to target" differences.
The effect of the individual weights on the allocation process is determined by their relative magnitude.
|
Setting |
Description |
| Major axis |
Weighting for the difference between the predicted location and the observed location of the target along the major axis. Weighting is a number between 0–100. |
| Minor axis |
Weighting for the difference between the predicted location and the observed location of the target along the minor axis. Weighting is a number between 0–100. |
| Range |
Weighting for the difference between the predicted location and the observed location for the range of the target. Weighting is a number between 0–100. |
| TS |
Weighting for the difference between the target strength of the last target in the open track and the observed target. Weighting is a number between 0–100. |
| Ping gap |
Weighting applied according to the number of pings between the last target in the open track and the observed target. Weighting is a number between 0–100. |
Notes:
- A weight value of zero means that the parameter does not contribute to the allocation process.
- 2D fish track detection will not use the Major and Minor axis weights.
See Fish tracking algorithms for more details of these calculations.
Fish track acceptance
These settings control how gaps are handled during fish track detection and which detected tracks are retained in the final results.
|
Setting |
Description |
|
Minimum number of single targets in a track |
The minimum number of single targets that a detected track must contain to be retained in the final results. This setting filters tracks after detection and does not affect the detection process. |
|
Minimum number of pings in track (pings) |
The minimum number of pings that a detected track must span to be retained in the final results. This setting filters tracks after detection and does not affect the detection process. |
|
Maximum gap between single targets (pings) |
The maximum number of consecutive pings without an accepted single target that may occur while a track remains open. Increasing this value can help connect tracks across longer gaps, particularly in high-ping-rate data. However, large values may increase detection time because more single targets can be considered for inclusion as a track is grown. |
Note: High-ping-rate data, including imaging sonar data, and data from slowly moving fish may produce tracks that span many pings. Larger values may be appropriate for these data.
Name and Notes page
Settings |
Description |
Name |
The unique name of the fish track detection configuration. In Echoview, name-uniqueness affects platforms, transducers, surfaces, scenes, region class, 3D curtains, 3D regions, lines, sticky notes, region detection configurations, raw variables and unavailable variables. Existing EV files, that contain objects with the same name, will open and affected objects will be assigned a unique name. Potential duplicate names are handled by preserving the source name and appending a bracketed number (e.g., (1)). Valid names for variables and dataflow objects consist of supported characters. These can include text and symbol character encoding and international characters. Commas, double quotes, the vertical bar and the Tab key press are unsupported characters. Where a name is used as the basis of an export file name, any unsupported file name characters are automatically removed. |
Notes |
A text box in which you can enter extra information about the fish track detection configuration. |
See also
Detecting fish tracks
Fish tracking algorithms
About beam geometry