decode_structured_light_pattern — Decode the camera images acquired with a structured light setup.
decode_structured_light_pattern decodes the camera images CameraImages that have been previously acquired with a structured light setup. The correspondence images and other intermediate results that are created by the decoding process are stored in the model StructuredLightModel and can be accessed afterwards using the operator get_structured_light_object.
In the following, the decoding process is explained in detail:
As mentioned in gen_structured_light_pattern the first purpose is to find out whether a pixel is in a region where a light stripe is reflected or where a dark stripe is reflected. To simplify this decision process the normalization images are used and a locally varying threshold is determined that is able to cope with objects of varying reflectance and lighting conditions. During the decoding of the acquired camera images all Gray code images are then compared with the previously calculated threshold. A pixel within the image is classified as bright if its gray value is greater or equal this threshold.
Furthermore, the pattern region is segmented during the decoding process. The segmentation is controlled by the parameter 'min_gray_difference' (see set_structured_light_model_param).
Assuming that n Gray code images have been processed, we get a n-bit binary code for each pixel. From this sequence the row and column coordinates up to of the monitor can be derived.
If the StructuredLightModel is a hybrid system consisting not only of Gray code images but also of phase shift images (see gen_structured_light_pattern), the next step is to decode the latter ones. The result is a subpixel-precise correspondence image between the monitor coordinates and the camera coordinates that contains the information which camera pixel observes which monitor pixel.
If the 'pattern_type' of the StructuredLightModel is set to 'single_stripe', the first step in the decoding process is to decide which single stripe shed its light on a camera pixel. The Gray code sequence and phase are then used to refine the position within the found single stripe.
In real world setups it may occur that the detected Gray code sequence of a pixel is wrong. This can then lead to values in the correspondence images which represent monitor rows or columns larger than the monitor width and height. To avoid these problems, the last step of the decoding process is to remove these values from the correspondence images.
This operator modifies the state of the following input parameter:
Acquired camera images.
Handle of the structured light model.
* Create the model create_structured_light_model ('deflectometry', StructuredLightModel) * Set the size of the monitor set_structured_light_model_param (StructuredLightModel, \ 'pattern_width', 1600) set_structured_light_model_param (StructuredLightModel, \ 'pattern_height', 1200) * Set the smallest width of the stripes in the pattern set_structured_light_model_param (StructuredLightModel, \ 'min_stripe_width', 8) * Generate the patterns to project gen_structured_light_pattern (PatternImages, StructuredLightModel) * Set the expected black/white contrast in the region of interest set_structured_light_model_param (StructuredLightModel, \ 'min_gray_difference', 70) * Decode the camera images decode_structured_light_pattern (CameraImages, StructuredLightModel) * Get the computed correspondences and defects get_structured_light_object (CorrespondenceImages, StructuredLightModel, \ 'correspondence_image') set_structured_light_model_param (StructuredLightModel, 'derivative_sigma', \ Sigma) get_structured_light_object (DefectImage, StructuredLightModel, \ 'defect_image')
The operator decode_structured_light_pattern returns the value 2 (H_MSG_TRUE) if the given parameters are valid. Otherwise, an exception will be raised.