| Summary: | Spectroscopy is widely used in various field, including in agriculture to determine the contamination of soil in order to produce the good quality of food and to avoid the excessive use of fertilizer, thereby minimize the impact on the environment. However, the commercial and common method of soil spectroscopy has some limitation such as bulky in size, costly and non-real-time system. In this study, a high-speed electronic data acquisition via machine learning on FPGA is implemented to efficiently monitor the macronutrients level in soil, which would offer economic benefit. Our focus is particularly on recognizing exact photon level absorbed by soil by applying photon count processing techniques to monitor the macronutrient in soil samples. The hardware architectures on FPGA feature a 16-bit Kogge Stone adder to process the input signals from the sensing module, LED light control system, time frame setting system and data synchronization via cloud for Internet of Thing (IoT) application. The proposed photon counting system has been demonstrated using visible range wavelength of 630, 550, and 470 nm, respectively. In addition, the input photon signal can be varied from 0 to 200 kHz and frame time period of 10 ms produces the optimum counting result with the percentage variation from 0% to maximum of 15% as compared to the actual counting from the signal generated by the function generator. Apart from that, a real-time system for IoT application has been successfully tested.
|
|---|