Over the past few months, we have been looking closely at the world of LiDAR scanning teamed with AI and machine learning, to see the uses it can have have in the world of agriculture.
What is LiDAR?
LiDAR has been in use since the 1960, with laser scanners being mounted on aeroplanes to provide aerial mapping. It wasn’t until the late 1980s, with the introduction of of commercially viable GPS systems, that LiDAR became a useful and accurate instrument for providing accurate geospatial data.
LiDAR is a remote sensing technology which uses the pulse from a laser to collect measurements of a specified area which can then be used to create 3D models and maps of both objects and environments. Although it is typically acknowledged that LiDAR stands for Light Detection and Ranging, this acronym was actually applied much later.
How does it work?
LiDAR uses light waves from a laser in a similar way that radar uses radio waves and sonar uses sound waves.
The time is calculated between the light leaving the laser and hitting the surface of an object, and the time it takes for the reflected light to travel back to the scanner. The lasers can fire around 1,000,000 pulses per second, and the data fed back to the sensors is used to build up a 3D image of the object or landscape that has been scanned.
This visualisation is also known as a ‘Point Cloud’.
What is it used for?
The most common use of LiDAR is for performing surveying and mapping tasks.
It is widely used for mapping built environments such as inside individual buildings, areas of buildings and road and rail networks. LiDAR Mapping uses a laser scanning system with an integrated Inertial Measurement Unit (IMU) and a Global Navigation Satellite System (GNSS) allowing each point in the Point Cloud to be georeferenced. This means that the maps give both absolute and relative positional accuracy.
The data gleaned can be used to map entire cities, pinpointing features such as landmarks, road and rail networks and areas of vegetation to name but a few.
LiDAR is also commonly used in environmental applications to detect things such as flood risk, carbon stocks in forestry and monitoring coastal erosion. The data can also be used to highlight any abnormalities in the landscape such as surface degradation, changes in slope gradients and vegetation growth.
LiDAR, AI and Machine Learning
There has been increased levels of use of LiDAR and AI in automation applications, and especially in autonomous vehicle navigation.
In products such as iDAR (https://aeye.ai/idar/) the AI is used to mimic how the human visual cortex evaluates the environment around a vehicle whilst driving, and how, as a human, we react to the driving conditions and hazards. The LiDAR generates an actual ‘view’ of the landscape, driving conditions and hazards which the vehicle can then react to once the AI has been added.
The volume and type of data that LiDAR can generate lends itself to being combined with AI and Machine learning in a push to further advance processes which are already being carried out.
What we have been looking at is how we can combine the uses of LiDAR, AI and Machine Learning to increase crop yield and optimise plant performance in a greenhouse environment.