ColorHistogramLabelMatch

../../_images/color_histogram_label_match.png

Compute similar region of image to specified histogram based on superpixels image.

Input Topic

  • ~input/histogram (jsk_recognition_msgs/ColorHistogram)

    Reference histogram.

  • ~input (sensor_msgs/Image)

    Input image. This image should be bgr8 or rgb8 image.

  • ~input/label (sensor_msgs/Image)

    Label of ~input image. Label image should be int32 image.

  • ~input/mask (sensor_msgs/Image)

    Mask image of ~input image. Only masked region is taken into account.

Publishing Topic

  • ~output/extracted_region (sensor_msgs/Image)

    Result of correlation computation as mask image.

  • ~output/coefficient_image (sensor_msgs/Image)

    Result of correlation computation as float image.

  • ~debug (sensor_msgs/Image)

    Debug image

Parameters

  • ~coefficient_method (Int, Default: 0)

    Method to compute coefficient.

    Choose from correlancy (0), chi_squared (1), intersect (2), bhattacharyya (3), EMD_Manhattan (4) and EMD_Euclid (5).

    This parameter can be changed by dynamic_reconfigure.

  • ~max_value (Int, Default: 255)

  • ~min_value (Int, Default: 0)

    Maximum and minimum index of histogram

    These parameters can be changed by dynamic_reconfigure.

  • ~masked_coefficient (Float, Default: 0.0)

    Value to fill masked region

    This parameter can be changed by dynamic_reconfigure.

  • ~threshold_method (Int, Default: 0)

    Method to binalize coefficient image.

    Choose from smaller_than (0), greater_than (1), otsu (2) and otsu_inv (3).

    This parameter can be changed by dynamic_reconfigure.

  • ~coef_threshold (Float, Default: 0.8)

    Threshold used in binalization.

    This parameter can be changed by dynamic_reconfigure.

  • ~use_mask (Default: false)

    Do not use mask image if this parameter is false.

Sample

roslaunch jsk_perception sample_color_histogram_label_match.launch