示例#1
0
class GraspActionServer:
    def __init__(self):
        self.server = actionlib.SimpleActionServer('object_pick', pickAction,
                                                   self.execute, False)
        self.result = pickResult()
        self.moveit = MoveIt()
        self.object_size = {
            "evergreen": [0.215, 0.05, 0.095],
            "tomatosoup": [0.1, 0.03, 0.13],
            "basetech": [0.16, 0.045, 0.17],
            "eraserbox": [0.155, 0.05, 0.055],
            "usbhub": [0.13, 0.07, 0.075]
        }
        self.server.start()

    def execute(self, goal):
        print('+++++++++++++++++++++++++++++++++++++++++++++++++')
        rospy.loginfo('Picking request received')
        self.moveit.close_fingers()
        x, y, z = goal.request.position
        z_max = goal.request.z_max
        rospy.loginfo("Received z_max from behav: {}".format(z_max))
        rospy.loginfo("Received z from behav: {}".format(z))
        success = self.moveit.clear_object(
            x, y, z, z_max, goal.request.yaw,
            self.object_size[goal.request.object_label])
        self.result = pickResult()  # Reset object array
        if success:
            self.server.set_succeeded(self.result)
        else:
            self.moveit.rtb()
            self.server.set_aborted(self.result)
示例#2
0
    	print "Deg: ", degrees(prediction)
    
        if np.array_equal(zero_image, image_depth) or np.array_equal(one_image,image_depth):
            print 'zero image'
            # uncomment to show binary depth image
            cv2.imshow("binary depth image" , image_depth[0])
        else:            
            #uncomment to show binary depth image
            cv2.imshow("binary depth image", image_depth[0])
            pass
            
	    image_depth = image_depth.reshape(1, 28, 28, 1)
        predictions.append(prediction)
        
        image_color = np.asarray([image_color])
        output_classes = model_classification.predict([image_color])[0]
        class_index = np.argmax(output_classes)
        
        #uncomment to show images
        cv2.imshow("color image", image_color[0])
        cv2.waitKey(1000)
            
    print "With orientation:", str(degrees(np.mean(predictions)))
    print "Box class:", class_index
 
    # Call the grasp function 
    # moveit.grasp(x, y, z, prediction, object_size[class_index])
    moveit.clear_object(x, y, z, z_max, prediction, object_size[class_index])

moveit.rtb()