powerful tools in many agriculture and livestock areas , enabling operators to create , modify , and share information .
Big Data Big Data refers to extremely large and complex data sets that cannot be analysed or processed using traditional methods . It requires advanced analytical tools and techniques to process these massive data sets .
Artificial Intelligence ( AI ) AI is a computational technology used to mimic human abilities to perceive their environment , learn , and make decisions . AI techniques play a key role in automation in the smart agriculture and food processing industry and can help accelerate the transition towards sustainable food systems . Generally , AI-based technologies can be used in real-time monitoring and decision-making processes .
Blockchain Blockchain technology can be defined as cryptographically secure distributed ledgers where the ledgers are distributed in a peer-topeer format among the entities , making it a decentralised system . Decentralised solutions are needed due to single points of failure , product irregularities , quality compromises , and loss of data present in conventional food supply chains . The adoption of blockchain in the agriculture and food supply chain , especially when combined with other technologies such as IoT , enables data interoperability , cost reduction , transparency , auditability , integrity and authenticity .
Imaging Technology for Smart Agriculture
There are still barriers and limitations to IR4.0 adoption in agriculture , such as not fully understanding the limitations of relevant technologies and data sources . Visual data is useful for AI model deployment where it can show important signs of plant or animal health and growth , such as by analysing its size and shape , biomass accumulation , photosynthetic activity , physiological status , water content and use , stress response , and biochemical composition . All this information can be obtained by using different types of images .
RGB images RGB images are obtained using a standard digital camera to capture images in red , green , and blue channels . This allows the capture of highresolution images of objects to assess traits like height , area , and growth patterns and also analyse colour-based traits such as colouration features , which can indicate health and stress levels . Together with deep learning , it has been used to automatically detect planthoppers in the paddy field and then classify them into four classes ( see Figure 1 ).
Hyperspectral image A hyperspectral image provides comprehensive data across a wide range of wavelengths , allowing for the detection of subtle differences in object composition and physiology . It is normally used in food safety , disease diagnosis , grading , classification , quality and composition analysis of animal products . As shown in Figure 2 , it can detect oil palm seedlings infected by disease at a very early stage i . e . in asymptomatic conditions .
Multispectral image A multispectral image is captured at specific wavelengths across the electromagnetic spectrum . Unlike RGB image , which captures only three bands , multispectral imaging collects data in typically three to 15 bands . This allows multispectral systems to provide valuable and affordable spectral information . Furthermore , it can be used to create a vegetation index . Figure 3 shows an example of the use of thermal images and vegetation index for disease detection .
3D image and Laser Scanning ( LiDAR ) Laser Scanning ( LiDAR ) uses laser pulses to create detailed 3D models of object structures . Meanwhile , Photogrammetry combines multiple 2D images taken from different angles to reconstruct 3D structures . The 3D image can be used to analyse complex traits like object architecture , biomass , and object structure ( Figure 4 ).
57