Abstract
Planet currently operates a fleet of over 200 satellites consisting of medium resolution monitoring SuperDove cubesats and high resolution tasking SkySats, with future very high resolution and hyperspectral tasking missions in development. In terms of image quality, this requires a processing pipeline extracting QA metrics, which is capable of handling these differences to ensure overall image quality stability.
Multiple teams within Planet are involved in our quality assurance concept, starting from system design and the manufacturing process, through payload maintenance, sensor and pixel calibration, and ending with a final quality assessment step independently from the processing pipeline and taking the perspective of the customer.
Planet’s Automated Quality Assurance (AQUA) team is responsible for this customer oriented final assessment step and focuses on geo-positional accuracy during the course of the last 4 years. The team has built a separate processing pipeline to extract quality metrics based on sampling from a millions-perday product stream. In this session, the AQUA team will present its separate production pipeline which continuously measures what we call key performance indicators (KPIs) of Planet’s fleet. This is purposely done using non-Planet resources and software, for example by employing independent reference imagery over a world spread set of test sites and then using the Commercial-Of-The-Shelf software to find checkpoints from Planet’s ortho products against the reference data set. The usage of independent resources in this process is intended and crucial in order to measure the product quality in such a way that the results can be reproduced by anyone outside Planet.
Fully Automated Earth Observation Imagery Quality Assurance
Planet currently operates a fleet of over 200 satellites consisting of medium resolution monitoring SuperDove cubesats and high resolution tasking SkySats, with future very high resolution and hyperspectral tasking missions in development. In terms of image quality, this requires a processing pipeline extracting QA metrics, which is capable of handling these differences to ensure overall image quality stability.
Multiple teams within Planet are involved in our quality assurance concept, starting from system design and the manufacturing process, through payload maintenance, sensor and pixel calibration, and ending with a final quality assessment step independently from the processing pipeline and taking the perspective of the customer.
Planet’s Automated Quality Assurance (AQUA) team is responsible for this customer oriented final assessment step and focuses on geo-positional accuracy during the course of the last 4 years. The team has built a separate processing pipeline to extract quality metrics based on sampling from a millions-perday product stream. In this session, the AQUA team will present its separate production pipeline which continuously measures what we call key performance indicators (KPIs) of Planet’s fleet. This is purposely done using non-Planet resources and software, for example by employing independent reference imagery over a world spread set of test sites and then using the Commercial-Of-The-Shelf software to find checkpoints from Planet’s ortho products against the reference data set. The usage of independent resources in this process is intended and crucial in order to measure the product quality in such a way that the results can be reproduced by anyone outside Planet.