This class provides functionality for finding lines in images. It will return candidate lines that meet the requirements specified by its properties.
Cognex.Vision ShareableBase
Cognex.Vision ChangedEventShareableBase
Cognex.Vision.LineMax.Implementation LineMaxOperatorBase
Cognex.Vision.LineMax LineMaxOperator
Namespace: Cognex.Vision.LineMax
Assembly: Cognex.Vision.LineMax.Net (in Cognex.Vision.LineMax.Net.dll) Version: 9.24.0.0
The LineMaxOperator type exposes the following members.
| Name | Description | |
|---|---|---|
| LineMaxOperator | Constructs a new instance of this class. |
| Name | Description | |
|---|---|---|
| Equals | Compares the equality of this object with another. (Overrides LineMaxOperatorBase Equals(Object).) | |
| Execute(IImage, IRegion) | Performs line finding on the supplied image. The inspection results are stored in a LineMaxResults object. | |
| Execute(IImageCollection, IRegionCollection) | Performs line finding on the supplied image. The inspection results are stored in a LineMaxResults object. | |
| Execute(IImage, IRegion, Image8Grey, Int32, Int32) | Performs line finding on the supplied image. The inspection results are stored in a LineMaxResults object. | |
| Execute(IImageCollection, IRegionCollection, Image8GreyCollection, Int32Collection, Int32Collection) | Performs line finding on the supplied image. The inspection results are stored in a LineMaxResults object. | |
| GetHashCode | Returns the hash code of this object. (Overrides LineMaxOperatorBase GetHashCode .) | |
| GetType | Gets the Type of the current instance. (Inherited from Object.) | |
| GetTypeMeta | Get the TypeMeta that describes the type of this object. | |
| ResumeAndRaiseChangedEvent | Re-enables raising of the ChangedEvent after SuspendChangedEvent has been called, and raises the ChangedEvent if the ChangedEventSuspended count is reduced to zero and any changes were made while events were suspended. Must be called once for each call to SuspendChangedEvent. | |
| SuspendChangedEvent | Temporarily suspends the raising of the ChangedEvent. May be called more than once, and a corresponding call to ResumeAndRaiseChangedEvent must be made for each call to SuspendChangedEvent. | |
| ToString | Returns a human readable string that represents the object. |
| Name | Description | |
|---|---|---|
| Equality | Compares the equality of two LineMaxOperator objects. | |
| Inequality | Compares the inequality of two LineMaxOperator objects. |
| Name | Description | |
|---|---|---|
| Changed | The actual event. |
Line finding is accomplished in two phases. The first phase, edge detection, finds edge points within the image(s). The second phase, line fitting, creates lines from roughly collinear subsets of the found edge points.
The first phase, edge detection, finds edge points within the image(s) using parameters defined in the LineMaxEdgeDetectionParams class. First, the input image is smoothed and decimated, using GradientKernelSizeInPixels. Gradient vectors are then computed for each pixel in the decimated image. Both the gradient vectors and the input image are projected over many caliper-like projection regions, using ProjectionLengthInPixels. A pixel is considered a candidate edge point if its projected gradient magnitude exceeds ContrastThreshold and if the ratio of its projected gradient magnitude and its intensity exceeds NormalizedContrastThreshold. Only edge points with gradient directions sufficiently consistent with ExpectedLineNormal can possibly constitute the expected lines at the angle ExpectedLineNormal (where the gradient direction consistency is determined by LineAngleTolerance and EdgeAngleTolerance). Therefore, only such edge points are included in the line fitting phase. Edge points with inconsistent gradient directions are excluded from the line fitting phase and placed into the UnconsideredOutliers collection.
The second phase, line fitting, creates lines from roughly collinear subsets of the found edge points, using parameters defined in the LineMaxOperatorBase class. Lines are evaluated using a RANSAC or an Exhaustive point selection methodology, depending on the FittingMode setting. Under RANSAC mode, the user may adjust Assurance to fit the desired result optimality and operation speed. Line fitting is performed by iteratively evaluating candidate lines created from the found edge points. During each iteration two points are randomly selected from the found edge points, a candidate line is modeled from the selected points, and the candidate line inliers are determined. An edge point is assigned to the inlier list of a candidate line only if it is in the proximity of the candidate line (determined by DistanceTolerance ) and if its gradient direction is consistent with the angle of the candidate line (determined by EdgeAngleTolerance ). Each edge point can only be assigned to the inlier list of a single line and each line is only allowed to contain at most one edge point from each projection region. When the RANSAC iterations reach the maximum, the inliers of the best found line candidate will be identified. Then the found line will be refined using a least squares regression fit to the inliers, and the set of inliers will be re-evaluated using the refined line. The refinement will be repeated a maximum of three times until the number of inliers ceases to further increase. Any edge points that are not inliers of the found line are placed into the ConsideredOutliers collection.