Skip to main content
Skip table of contents

Qimera CUBE Settings Dialog

How to Start

  • CUBE sub-menu of the Dynamic Grid context menu option in the Project Layer dock.

What it Does

This dialog provides editing of key CUBE processing parameters. 

General Description

The uncertainty associated with the CUBE nodes is the total vertical uncertainty as built from the soundings which:  (1) fall within the capture radius of the node, and (2) are consistent with the primary node hypothesis. Primary node hypotheses are determined from the Disambiguation method defined at the time of CUBE grid building (Most Soundings, Nearest Neighbors, etc.). The uncertainties associated with the soundings are both Horizontal and Vertical, based on all the individual uncertainties associated with the systems that are associated with a MBES configuration. 

Setting

Description

Distance Scale

Scale on predicted or estimated depth for how far out to accept data.  (unitless; typically 5% for hydrography but can be greater for geological mapping in flat areas with sparse data)

Distance Min.

Only accept data with this radius.

Resolution Algorithm

When running the CUBE algorithm, sometimes multiple hypotheses are generated. The algorithm must make a best guess on which hypothesis is correct using a Hypothesis Resolution Algorithm. The following algorithms are supported:

  • Num. Samples – The hypothesis containing the most soundings will be chosen.

  • Neighborhood – The algorithm searches all the neighbors of the cell looking for a cell with only one hypothesis. If a cell is found with one hypothesis, this cell is used as a guide in choosing a hypothesis; otherwise the Num. Samples algorithm is used.

  • Num. Samples + Neighborhood – A combination of the two previous algorithms.

  • Predicted Grid – The algorithm computes the median grid of the input soundings as uses that grid as a guide in choosing the hypothesis.

Estimate Offset

Threshold for significant offset from current estimate to warrant an intervention.

Horizontal Error Scale

This parameter is used to control the exponential growth of a soundings vertical uncertainty as a function of distance from the node.

Distance Exponent

Exponent on distance for variance scale.

Queue Length

Length of median pre-filter sort queue.

Quotient Limit

Outlier quotient upper allowable limit.

Discount Factor

Discount factor for evolution noise variance.

Bayes Factor Threshold

Bayes factor threshold for either a single estimate, or the worst-case recent sequence to warrant an intervention.

Run Length Threshold

Run-length threshold for worst-case recent sequence to indicate a drift failure and hence to warrant an intervention.

CUBE Capture Distance specifies how close a sounding must be in x/y distance to where the seafloor is being estimated. When building a dynamic grid the estimation point is the center of each cell in the grid. Thus you should never choose a distance which is smaller than the diagonal length of a cell or the grid might not get a value contribution from all samples in the incoming data. Generally speaking a diameter that is 1.25 * the cell size is a good option. The Distance Scale option adjusts the capture radius based on depth. For example if you choose 5% and the water depth was 25 meters for a particular incoming sounding then it would have a capture radius of 1.25m. This means it would contribute a value to any cells whose center was no more then 1.25 meters away from the measurement. The Distance Minimum is exactly that - every sounding is assigned exactly the same capture diameter as that which is given. 

Internally, the CUBE algorithm has no concept of cells. A sounding may contribute to many different hypotheses depending on how far away the sounding is away from the hypothesis center in the XY plane. The Capture Distance option controls how far the sounding can be to contribute to a hypothesis and the algorithm assumes that deeper soundings will be less accurate. If the distance from the hypothesis center to the sounding is less than the sounding height multiplied by the Capture Distance percentage, the sounding will be included in the hypothesis. Smaller Capture Distance values result in less soundings contributing to a hypothesis and larger values result in more soundings contributing to a hypothesis. If the Capture Distance is too small, holes may appear in the data. In some cases, this Capture Distance does not make sense so to turn the algorithm off, enter 100 as the Capture Distance value.

Return To:  CUBE Dialogs

Return To:  Dynamic Surface Context Menu

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.