clear_CHARLES
Autonomous onboard data processing for passive payloads.
SMART, AUTONOMOUS ONBOARD DATA PROCESSING.
High-Quality Data, Delivered Onboard.
clear_CHARLES is a state-of-the-art onboard data processing solution, specifically engineered for passive payloads.
Leveraging sophisticated algorithms, clear_CHARLES enables real-time cloud detection, precise object tracking, and sensitive change detection.
By performing these critical tasks onboard, clear_CHARLES significantly reduces ground data processing requirements and accelerates time-to-insight.
Forget Ground Delays,
Get Ready-to-Use Data.
clear_CHARLES operates on a modular basis, allowing for customization and integration with various passive payload systems.
Data acquisition is the first step, where clear_CHARLES receives raw image data from the payload's sensors. This data is then preprocessed to reduce noise, normalize the images, and correct for geometric distortions.
Feature extraction follows, where relevant features such as edges, textures, and color information are extracted from the images. These features are then used by AIKO machine learning models to classify pixels or objects within the images, identifying clouds, objects of interest, or changes.
Data filtering and compression are essential for efficient transmission. Irrelevant or redundant data is filtered out, and the remaining data is compressed to reduce file size.
Finally, the processed data, including segmented images, detected objects, or change maps, is prepared for transmission to ground or for further onboard analysis
Maximize Data Quality and Transmission Efficiency.
Passive payloads rely on external sources of energy, such as sunlight, for operation. clear_CHARLES offers significant advantages for such payloads:
Reduced Power Consumption: Onboard processing can reduce the amount of data transmitted to ground, conserving precious battery power.
Real-time Decision Making: By analyzing data in real-time, passive payloads can autonomously adjust their operations based on changing conditions, such as cloud cover or target movement.
Increased Mission Efficiency: Identifying and focusing on areas of interest before downlinking data can optimize mission objectives and reduce operational costs.
Enhanced Data Quality: Onboard processing can filter out irrelevant or noisy data, improving the quality of information transmitted to ground.
Real-time Results, Real-World Impact.
AIKO's cloudy_CHARLES technology demonstrated its capabilities in a successful in-orbit demonstration with Satellogic in 2019. The algorithm was specifically fine-tuned for the mission requirements and optimized for the satellite's onboard processing hardware. Key achievements include:
Reduced Memory Footprint: The compiled application achieved a compact size of 21MB, minimizing onboard storage requirements.
Optimized Performance: The algorithm was tailored for the target processing computer, ensuring efficient execution and low inference times.
Successful Deployment: cloudy_CHARLES was successfully integrated into the satellite's operating system and processed images from the onboard camera.
Real-time Performance: The algorithm demonstrated real-time capabilities, with inference times of approximately 2 seconds per frame.
HERITAGE
clear_CHARLES is the development of AIKO’s cloud segmentation technology, cloudy_CHARLES. Since its inception, cloudy_CHARLES has undergone rigorous development and testing, demonstrating its ability to deliver exceptional performance in demanding space environments.
Key milestones include:
Early development: initial proof-of-concept testing in 2018 on consumer-grade hardware laid the foundation for cloudy_CHARLES.
Adaptability: successful testing on space-graded hardware platforms in 2019 showcased the algorithm's versatility and robustness.
In-Orbit demonstration: the technology's effectiveness was demonstrated in a groundbreaking in-orbit demonstration in 2019.
Continuous refinement: AIKO has consistently improved cloudy_CHARLES, enhancing its accuracy, efficiency, and compatibility with various hardware platforms.
From first IOD, AIKO’s has also enhanced new features and machine learning models, by creating the current clear_CHARLES solution.
SYSTEM REQUIREMENTS
- OPERATIVE SYSTEM - Linux-based OSs CPU COMPATIBILITY - ARM 32bit/64bit Cortex-A series, X86_64 processors
- HW ACCELERATORS: Nvidia GPUs; Xilinx and Microchip FPGAs; Intel Myriad VPU; Google Edge TPU
- RAM and DISK SPACE: depending on payload objective