Research and Development in Robotic Systems

Operating and Evaluation Software for Sewer Inspection

The Robotic Systems Business Unit's development of the inspection systems for Emscher sewer system also entailed the development of all the system components for data processing.

© Fraunhofer IFF

Inspection and Cleaning Robots for Sewer Systems: Data Processing

Software Architecture

The unique challenges service robotics presents to software resulted in the development of a novel, service-oriented software architecture They facilitate efficiently, reliably and fully recording, analyzing, archiving and visualizing sensor data and the results of processing. Ultimately, this provides an advanced framework for inspection robots, which may also be adapted for many other service robots.

Operating and Evaluation Software

The inspection systems for sewer systems generate extensive and partly abstract data during their use. A package of operating and evaluation software (BASW) was developed to manage these extremely large quantities of data. BASW not only provides operators the world's first clear 3D representation of inspected sewer lines but also extreme close-ups and special aids to precisely evaluate, quantify and measure potential damage. The BASW was designed for use as a tool for post-viewing and analysis of stored datasets and as a remote operating environment during inspection.

© Fraunhofer IFF
Operating and evaluation software.

Position Detection and Visual Odometry

Constant knowledge of position and orientation in an environment is key for mobile service robots in general and inspection systems for sewer systems in particular. Hence, the development of the software for the systems entailed the development of a solution that visually determines the floating inspection system's (SEK) position without markers and the development of a larger, wheel-guided system (SVM).

Developed to product maturity, the solution precisely localizes the equipment inside underground, partially filled concrete pipes in all six degrees of freedom and serves as the basis for registering measured data and visualizing the sewer line. For instance, merely knowing the exact position enables compiling measured data in the form of high resolution camera images into an image of the entire sewer line like a mosaic.

Data Evaluation and Crack Detection in Concrete Surfaces

The mass of data from automatic inspection cannot be examined by the operator alone: support is necessary. To display potential damage on the inspected sewer walls, the inspection systems take seven or nine high-resolution camera pictures per second. The multitude of images would make purely manual inspection extremely time or labor consuming. Therefore, an automatic image analysis system was developed, which automatically detects and evaluates cracks with a width of 0.5 mm or more even under the adverse conditions in a sewer line, displays them to the user and logs them. The diversity of forms of cracks, the concrete surface, the specific light conditions, contamination and surface moisture necessitate a complex system of algorithms, which has not been commercially available until now. Other software modules support the user with measuring joint width or detecting deposits.

Image processing for the inspection with service robots: Examples from sewer inspection cover VR visualization with operation and navigation in large amounts of data, automatic crack detection, automatic joint detection and measurement as well as visual tracking for positioning.