Active Safety Architecture Uses Cameras for Safer Driving
In an earlier article, we spoke about the reasons Mindtree joined AUTOSAR to improve automobile safety while expanding infotainment capabilities. As an active AUTOSAR partner, Mindtree is at the forefront of including Active Safety capabilities that address key risk factors. For example, our Road Surface Monitor detects conditions like potholes (sharp depressions in road surfaces) and speed bumps. This enables drivers to take the necessary actions to avoid potholes and reduce the impact of speed bumps, thereby facilitating a safer driving experience.
To enable our Active Safety solutions, we use cameras to augment human visual capabilities. By applying advanced image processing techniques to videos and images, we derive useful information that can produce distinctions based on defined criteria, such as distinguishing road bumps or depressions from flat surfaces. These distinctions are based on a series of algorithm steps:
- Image enhancement by median filtering.
- Intensity-based adaptive threshold to convert image into binary image, with black image representing object of interest.
- Morphological erosion/dilation of image to remove small clusters in binary images.
- Apply connected component label and chain coding technique to count number of objects, area and perimeter of each object.
- Contour-based algorithm for reading the object properties and identifying it as a pothole or flat road surface.
It’s essential to alert drivers to both potholes and speed bumps to protect their cars and passengers. According to the U.S. National Highway Traffic Safety Administration, almost 4% of auto accidents were caused by poor road conditions, including heavy rain, oil spills, and road hazards. With night driving, headlights don’t always reveal unexpected road conditions in time, thus adding to frequency of accidents. Millions of people face these dangerous issues every day. A precautionary mechanism that proactively alerts drivers to bad road conditions can significantly enhance vehicle and driver safety.
For example, with real-time monitoring of road conditions, an alert system can detect potholes or other conditions before they are encountered. The road condition alert system not only improves safety, but also reduces the wear and tear of vehicles caused by driving over bad roads with potholes or other impediments. This monitoring is accomplished using small sensor cameras built into the vehicle during production or added to the vehicle as an aftermarket product.
To reduce accidents and other problems caused by speed bumps, Mindtree has developed a system using cameras with enhanced imaging to assist drivers with proactive alerts that monitor road issues in real time. Similar to road condition alerts, the system provides a warning that a speed bump or other anomaly has been detected in the road, so drivers can take appropriate action.
To differentiate between flat road surface and indented potholes, the camera captures the following parameters – standard deviation, circularity and average width. These data points are compared with a predefined area perimeter of the pattern of pothole, patch contour and cracks contour and the system concludes if they are real potholes or some other object on the road. See figure 1 below:
Figure 1: Mindtree uses OpenCV-based algorithms to distinguish impediments on the flat road surface and alert the driver. Once alerted, drivers can slow down to avoid the bad road conditions.
At Mindtree, we’re continually looking for ways to improve automobile safety. Our AUTOSAR partnership has helped us accelerate the development life cycles of our Active Safety systems for road condition and driver drowsiness alerts. The AUTOSAR standard interfaces and services are configured to make the solution scalable; and application-level abstraction has allowed us to reuse the interfaces for future projects.
In the next article, we’ll talk about the cool technology being used by Mindtree to alert drivers when they show signs of drowsiness.