Object detection device and model pattern evaluation device
A model pattern evaluation device for a robotic system comprises a processor. The processor is configured to: receive an image in which a target object is represented captured by a camera attached to a movable robotic component of the robotic system; calculate, for detecting the target object from the image, an evaluation value representing a geometric distribution of a plurality of features in a model pattern, the model pattern corresponding to the target object; set, among the plurality of features, a first feature for detecting a position in a specific direction of the target object in the image or a second feature for detecting an angle in a rotational direction centered about a predetermined point of the target object in the image as a feature of interest based on the evaluation value; and generate a detection result based on the first or second features to control the movable robotic component.
1. A model pattern evaluation device for a robotic system, comprising:
a processor configured to:
receive an image in which a target object is represented captured by a camera attached to a movable robotic component of the robotic system;
calculate, for detecting the target object from the image, an evaluation value representing a geometric distribution of a plurality of features in a model pattern, said model pattern corresponding to the target object;
set, among the plurality of features, a first feature for detecting a position in a specific direction of the target object in the image or a second feature detecting an angle in a rotational direction centered about a predetermined point of the target object in the image as a feature of interest based on the evaluation value; and
generate a detection result based on the first feature or the second feature to control the movable robotic component of the robotic system.
2. The model pattern evaluation device according to claim 1 , wherein the processor is configured to:
specify a direction in which a spread of a distribution of positions of the features is maximized as the specific direction based on the evaluation value; and
set, among the plurality of features, a feature which is distributed in a direction different from the specific direction as the first feature.
3. The model pattern evaluation device according to claim 1 , wherein the processor is configured to acquire a normal vector relative to the direction in which adjacent features are aligned for each of the plurality of features, calculates a covariance matrix representing the distribution of the normal vector as the evaluation value, and when the ratio of the smaller eigenvalue to the larger eigenvalue of the two eigenvalues of the covariance matrix is equal to or less than a first threshold value, sets a direction of the eigenvector corresponding to the larger eigenvalue as the specific direction.
4. The model pattern evaluation device according to claim 1 , wherein the processor is configured to:
specify that a part of the features is distributed along the rotation direction centered about the predetermined point based on the evaluation value, and
set a feature distributed in a direction different from the rotational direction as the feature of interest.
5. The model pattern evaluation device according to claim 4 , wherein the processor is configured to acquire a normal relative to a direction in which adjacent features are aligned for each of the plurality of features, at least a part of the normals intersecting at the predetermined point, set the ratio of the number of normals which do not intersect at the predetermined point to the total number of the features as the evaluation value, and specify that the part of the features is distributed along the rotation direction centered about the predetermined point when the ratio is equal to or greater than a second threshold value.
6. The model pattern evaluation device according to claim 5 , wherein the processor is configured to set a feature for which the normal does not intersect at the predetermined point as the feature of interest.