On the heels of announcing it’s expanding tests into Florida, Waymo has released the Waymo Open Dataset for autonomous vehicle researchers. The Waymo Open Dataset, available for free, covers a wide variety of environments, from dense urban centers to suburban landscapes. It also includes data collected during day and night, at dawn and dusk, in sunshine and rain.
The Alphabet Inc. subsidiary believes the Waymo Open Dataset is one of the “largest, richest and most diverse self-driving datasets ever released for research.” The Waymo Open Dataset website notes “this data is licensed for non-commercial use.” It continues, “while this dataset is not reflective of the full capabilities of our sensor system and is only a fraction of the data on which Waymo’s self-driving system is trained, we believe that for research purposes this large, diverse and high-quality dataset is extremely valuable.”
Here is a breakdown of the Waymo Open Dataset:
Size and coverage: This release contains data from 1,000 driving segments. Each segment captures 20 seconds of continuous driving, corresponding to 200,000 frames at 10 Hz per sensor. Such continuous footage gives researchers the opportunity to develop models to track and predict the behavior of other road users.
Diverse driving environments: This dataset covers dense urban and suburban environments across Phoenix, AZ, Kirkland, WA, Mountain View, CA and San Francisco, capturing a wide spectrum of driving conditions.
High-resolution, 360° view: Each segment contains sensor data from five high-resolution Waymo lidars and five front-and-side-facing cameras.
Dense labeling: The dataset includes lidar frames and images with vehicles, pedestrians, cyclists, and signage carefully labeled, capturing a total of 12 million 3D labels and 1.2 million 2D labels.
Camera-lidar synchronization: Waymo has been working on 3D perception models that fuse data from multiple cameras and lidar.
Waymo is universally viewed as the most advanced company in the autonomous vehicle industry. Other self-driving companies such as Lyft and Argo AI have also recently open-sourced self-driving datasets. These releases could be viewed as an admission that building Level 5 self-driving cars is much harder than previous thought. But Waymo head of research Drago Anguelov said that’s not the case.
“It’s not an admission, in any way, that we have problems solving these issues,” he said. “We felt, and it was not just us … that the field was currently hampered by a lack of suitable datasets.”
“The fact that Waymo and other folks are releasing the dataset is not so much about, ‘Hey, the problem is so hard, let’s just pool our data’; it’s about how we empower the research community for whom it’s really hard to get access to datasets like this,” added Waymo Product Lead Vijaysai Patnaik.
The Waymo Open Dataset, the company said, has the potential to help researchers make advances in 2D and 3D perception, and progress on areas such as domain adaptation, scene understanding and behavior prediction.
“We hope that the research community will generate more exciting directions with our data that will not only help to make self-driving vehicles more capable, but also impact other related fields and applications, such as computer vision and robotics,” Waymo said in a statement.