Essess has invested millions of dollars over the past 4 years developing the leading drive-by thermal imaging capability in the market today. Our imaging vehicles are equipped with technology to enable our unique data collection capabilities:
Multi-spectral infrared imaging of structures
- Long wave infrared (LWIR) radiometric cameras
- Near-infrared (NIR) high dynamic range cameras
- NIR scene illumination for rural and poorly lit suburban regions
- Capture thermal signatures of structures
- Discovery of building facades and background removal using computer vision and machine learning engines
- Camera housing offers 70 degree vertical field of view and full width horizontal field of view of structure due to vehicle motion
Automated building detection within property boundaries
- Rotating laser array light detection and ranging (LIDAR) sensor captures ranging and reflectance even from large standoff distances
- Buildings are isolated from the scene using 3D LIDAR point clouds
- Ranging capability allows structures to be bounded within property lines and relevant locations
- Mapping grade Global Positioning Systems (GPS) and support filtering algorithms ensure that location of structures and properties are highly accurate
- Collected data used in simultaneous localization and mapping (SLAM), which allows Essess to supplement the GPS data captured and more accurately correlate each image to the relevant home
Highly reliable on-board data capture and diagnostics system
- Performs over wide range of seasonal temperatures, down to at least -20F (-30C)
- High mast enables operation in a variety of regions, including short standoff distances with 3-4 story buildings
- On-board data validation and recording software and hardware
- Real-time diagnostic and quality control provided by LTE cell network streaming to Essess headquarters
Combined, these hardware and software capabilities constitute the single most effective way to capture heat loss and building envelope data via drive-by thermal imaging. Each camera captures data in a video format, meaning that Essess generates hundreds of thousands of images comprising over 2 terabytes of data each night. The LIDAR sensors enable Essess to generate a 3D map of the physical environment and map buildings to parcels in a highly accurate manner. Our unique hardware setup allows us to capture vast amounts of data and subsequently process that data in a very efficient process.
The below diagram provides a very general view of some of the key steps in our data processing pipeline:
High Speed Storage – Each imaging vehicle captures several terabytes of data per night which is stored in the Customized Vehicle Data Storage System.
File Expansion & Compression – The compressed data is extracted from the hard drives in order to begin processing and analysis. Each night, Essess captures terabytes of data.
Vertical Stitching – The vehicle is equipped with two LWIR cameras on each side of the imaging device and each camera captures a part of the scene as the vehicle drives-by. In order to get a robust, vertical image of the scene, data streams from the two cameras are stitched together using proprietary Essess algorithms.
Geo-location – All of the data from the GPS units is analyzed and then combined with LIDAR information to adjust for any external noise or loss of satellite signal.
Building Matching – Once the GPS data is processed and analyzed, it is matched up with the relevant thermal images for each home imaged.
Horizontal Panorama – As the data is captured frame by frame, there maybe tens or hundreds of images with each image showing a small part of the entire scene. In order to get a seamless panorama of an entire building, the frames must be stitched together.
Energy Scoring – Once the images are extracted, vertically stitched, correlated with the relevant address, and horizontally grouped, they are analyzed to convert the thermal reading into an energy score. This energy score is relevant to each data set and allows for one home to be compared to a different home within the same dataset.
Leak Detection – The images are also analyzed for potential building envelope leaks.
Low Speed Storage – All of the raw data is then placed in low-speed storage.
Database – The analyzed and processed data is stored into a database. Essess customers can then access this data using front-web applications layered on top of the database.