基于ROS 目标检测与识别-Object Detection and Recognition

Object recognition has an important role in robotics. It is the process of identifying an object from camera images and finding its location. Using this, a robot can pick an object from the workspace and place it at another location.This chapter will be useful for those who want to prototype a solution for a vision-related task. We are going to look at some popular ROS packages to perform object detection and recognition in 2D and 3D. We are not digging more into the theoretical aspects, but you
may see short notes about the algorithm while we discuss their applications.
**Getting started with object detection and
recognition**
So what’s the main difference between detection and recognition? Consider face detection and face recognition. In face detection, the algorithm tries to detect a face from an image,but in recognition, the algorithm can also state information about whose face is detected. It may be the person’s name, gender, or something else.
Similarly, object detection involves the detection of a class of object and recognition performs the next level of classification, which tells which us the name of the object. There is a vast number of applications that use object detection and recognition techniques. Here is a popular application that is going to be used in Amazon warehouses:
Amazon is planning to automate the picking and placing of objects from the shelves inside their warehouses. To retrieve objects from the shelves, they are planning to deploy robotic arms such as the one shown in the previous image. Whenever the robot gets an order to
retrieve a specific object and place it in a basket, it should identify the position of object first, right? So how does the robot understand the object position? It should need some kind of 3D sensor, right? And also, on the software side, it should have some object recognition
algorithm for recognizing each object. The robot will get the object coordinates only after the recognition. The detected coordinates will be relative to the vision sensors, which have to transform into robot end-effector coordinates of the tip of the robot to reach the object position. After reaching the object position, what should be the robot do? It should grasp the object and place it in the basket, right? The task looks simple, doesn’t it? But it’s not as simple as we think. Here is the coordinate system of a robotic arm, end effector, Kinect, and the object:
Amazon organizes a challenge called the Amazon Picking Challenge, which was first
conducted as a part of ICRA 2015 , and in 2016, it was conducted along with Robocup. The challenge was all about solving the pick-and-place problem we just discussed. In effect, the object recognition and detection tasks have immense scope in the industry, not only in Amazon but also in areas such as agriculture, defense, and space.