Posted on

Controlling and ensuring safety of the Lego robot arm: the computer vision challenge

As the Lego robot arm executes a mission, information about the environment is required. Beside executing missions correctly, our goal is to detect any kind of danger caused by the robot. The goal is twofold: the robot has to know when to execute a mission i.e. the object to be moved is at the right place. Second, it has to stop when some dangerous situation happens, for example a human is present near to the robot.

We are building the monitoring infrastructure of the robot arm based on computer vision technologies. OpenCV helps us detecting and tracking the movements of the robot In case of a moving robot, no other moving objects should be there. In addition, computer vision will detect if the object to be transported is in the proper place to handle by the robot.

First time we built only a robot arm with limited functionality. According to the experiences, we have totally rebuilt it.

Rebuilding the robot was a big step forward for the project’s computer vision goals. Now, we are able to detect the orientation and also the movement of the arm, without markers.

The new concept is to put on some Lego element in a combination to form a larger component with distinctive shape and colour. The camera observes the whole loading area from the top, and searches for the elements.

Using the same camera frames, we can detect the orientation of the gripper and find the cargos and the train.

For the gripper we needed a marker and that Lego element which we talked about before. The marker is directly connected to the gripper’s motor, so it is moving with the gripper during the rotation and other movements. The orientation is compared to the arm’s orientation so it won’t change during the movement of the arm, only the rotation influences its settings. The marker is a black circle, therefore we replaced the color detection with circle detection for this case.

Cargo has distinctive color.

In the following picture the output of the various steps of the process is depicted.

Output of the image processing steps

In the following we just sketch the working of the detection algorithms. Transforming the picture of the camera to HSV (Hue-Saturation-Value) representation. This will serve as a base representation for further processing. The next step is to decompose the picture according to the information we are looking for. In order to ease the tasks of the further processing, the picture is cut into pieces: The Lego arm, the gripper and also the object to be moved will be in different pieces of the picture.

Detecting the various objects of interest, we need to assess the color and the size of the objects in the picture: this is assessed at the next phase of the processing.

Edge detection algorithms search for the contour of the objects. Pattern matching algorithms try to find rectangles in the picture.

Numerical filters than used to sort the found objects (rectangles) according to their size. From the filtered objects, some special heuristics filter those object which are likely to be the searched object, namely the arm, the gripper and the load to be moved.

The movement of the arm is traced by reducing the problem to finding the moving rectangles in the filtered picture. Computing averages and tracing the middle point of the objects provide quite precise results.

So, as you might see, many algorithms work on the control and safety assurance of the Lego robot arm. Despite its complexity, it works well in practice!

Posted on

Computer vision based safety-system: how to get the information

The system we described was originally operated by distributed units, called masters. These masters got the local information about occupancy through a special circuit integrated into the board. However, network problems often caused the error in the safety-logic, so we decided to introduce an additional layer of safety based on computer vision and complex event processing.

Now, we will give some details about the application of computer vision for recognizing the trains and their location.

The safety logic deployed to the embedded controllers have binary information of the trains, namely if a train is on a section of the system, we detect it. There is no information about the direction of travel, and speed. Because these limitations, the information of the safety logic is rather limited.

Because the logic itself cannot determine the direction, it must consider the worst-case scenarios. This causes deadlocks, and unnecessary stops. This is a price we pay for safety.

The previously mentioned solution operates in a distributed manner. It’s safe, it’s reliable, it’s formally verified. If everything works correctly.

So we decided to implement the runtime verification of the local components and we integrated the system level monitor based on computer vision. We show the later now in details.

Our monitoring solution is a computer vision based one, using the open-source OpenCV framework. OpenCV is a very extensive library of optimized image processing, machine learning algorithms, ideal for quick development of computer vision based applications. You are not worried about the performance and programming complexity.

There are other solutions maybe with better performance, however as OpenCV is open-source and there is a huge community behind it, our decision was straightforward.

This is an example marker we use on the top of the trains. There are three markers: red, green, and a blue one.

Our needs were pretty simple: identify the trains, and determine their positions. Circular patterns are great for this kind of computer vision tasks, because if you rotate a circle, nothing happens, therefore you don’t have additional complexity.

So we decided to use markers to make our task easier!

Many of the people reading this article may think about the Hough circle algorithm, which can find circular patterns. The problem with this algorithm is it’s genericity: our board may contain many circles, not just only the train. We needed an error prone algorithm, which can match a specific pattern, if only a partial circle is visible.

What we can do is use some math. Instead of traditional pattern matching, we can turn this into a math problem. Our pattern is very static. By static, I mean the circle pattern doesn’t vary by size. Because of this property, we can create a very specific matcher, using convolution. Convolution is basically two math functions, and we apply one function on the other. The resulting function is the combination of the two. Although convolution is quite difficult to achieve, but if we transfer our image to the frequency domain, the convolution basically becomes a multiplication which is easy to do.

Let’s see an example what is happening:

  1. We create a pattern image, with specific values. These values can be: 0, if we are not interested what’s really there; 1, if we want this area white; -1, if we want this area black.
  2. Convert this pattern image to frequency domain.
  3. Read the image from camera.
  4. Convert this image to the frequency domain.
  5. Multiply the two spectra.
  6. Convert the multiplied image back.

The pattern

This is a pattern, where green has value 0, white has value 1, and black has value -1

The image from the camera looking down the MoDeS3 board

The camera image’s, and the pattern’s convolution result

This is not a pitch black image, if you look closely, you can see the bright points, which marks the points where markers found

Now we have a weird-looking image, where a brighter spot means a bigger match between an image, and the pattern. On this image, we can use a simple threshold, and get a binary image, where it is really trivial to find the brightest points.

We are not saying this is the most efficient algorithm for this solution, but it’s really robust, and precise. The precision is in the millimetre range, and it’s robustness can be described as this solution does not make false detection. It might not detect valid points for a small time period, but we haven’t seen false reading, not even in a 8-hour-long session. On the other side, Complex Event Processing can solve the problem when false values are observed for a small time interval.

So what’s after the detection of the circle pattern? There is a color ID inside the two circle patterns, and this color identifies the train. What we do is search for points, where the distance of these two points exactly matches the distance in the real world. If we find a pair, we can be sure it’s a train marker. After idetifying all the visible trains, we convert the data to JSON, and publish it to the MQTT broker.

Our approach may seem a little non-standard, but it’s proven it’s reliability, and after all, that’s what matters for us.