Mapping Agricultural Soil in Greenhouse Using an Autonomous Low-Cost Robot and Precise Monitoring

Our work is focused on developing an autonomous robot to monitor greenhouses and large fields. This system designed operate autonomously extract useful information from the plants based precise GPS localization. The proposed RGB camera for plant detection a multispectral extracting different special bands processing, embedded architecture integrating Nvidia Jetson Nano, which allows us perform required processing. uses multi-sensor fusion manage two parts of algorithm. Therefore, algorithm was partitioned CPU-GPU architecture. process each image in 1.94 s sequential implementation approach followed our Hardware/Software Co-Design study propose optimal implementation. experiments were conducted tomato farm, showed that we can images real time. parallel 36 ms allowing satisfy real-time constraints 5 images/s. On laptop, have total processing time 604 9 In this context, obtained acceleration factor 66 laptop 54 energy consumption evaluation prototyped consumes power between 4 W 8 W. For raison, case, opted low-cost Nano.