This project introduces a novel, low-cost, real-time soil analysis device that integrates multi-sensor technology, AI-driven analytics, and IoT connectivity to enhance precision farming in Ghana and Zambia. While existing soil testing methods rely on laboratory-based analysis or expensive commercial sensors, this project focuses on developing an affordable, portable, and AI-powered solution tailored for smallholder farmers. The device will provide instant soil health assessments and generate personalized crop and fertilizer recommendations, making data-driven precision farming accessible to resource-limited farmers.

Funders: The Ministry of Environment, Science and Technology (MEST), Ghana, and the National Science and Technology Council (NSTC), Zambia, under the Zambia–Ghana Bilateral Joint Call for Collaborative Research Proposals in Emerging Technologies and Development, Space Science & Technology.

Duration: September 2025 - August 2028

Collaborating Institutions: 

• CSIR-Institute for Scientific and Technological Information (CSIR-INSTI)

• University of Zambia (UNZA)

• Bolgatanga Technical University (BTU)

• Golden Valley Agricultural Research Trust (GART)

• National Remote Sensing Center (NRSC)

• The seminar focused on How to Write a Research Paper: A Practical Case Study, using draft manuscripts prepared by LSACROFT postgraduate students Sarah Annang and Latifa Yakubu from the Bolgatanga dataset.

The session covered:

• 1. Structure and key components of a research paper

  2. Effective presentation and discussion of research findings

  3. Common weaknesses in research manuscripts and how to address them

  4. Preparing manuscripts for journal submission

The session was presented by Dr. Pennante Naa Ayikailey Bruce-Vanderpuije, Senior Research Scientist, and held on 3rd September 2026 at CSIR–INSTI, with both in-person and virtual participation. link to the meeting is below:

• How to Write a Research Paper: Practical Guidance Using LSACROFT Case Studies

• The LSACROFT Project Dataset Review Meeting, bringing together research teams from Ghana and Zambia to discuss the type of data gathered during the Cowpea & Maize Variety Trial at the Bolgatanga Experimental Farm was held on Friday, 14th August 2026 via Google Meet. A link to the meeting is below:

  • LSACROFT Project Dataset Review Meeting

• Three papers from the LSACROFT project presented at the 49th ICT and Electronics Convention (MIPRO 2026), have been published in IEEE XPLORE.

  1. A Low-Cost Hardware–Mobile Platform for Comprehensive Soil Monitoring, Crop Phenotype, and Image Dataset Acquisition.

  2. Survey and Design Framework: Comparative Evaluation of Edge Computing Platforms for Low-Cost AI-Enabled Real-Time Soil Monitoring, Dataset Collection, Crop Recommendation, and Fertilizer Optimization in African Settings.

  3. A Comprehensive Survey of Low-Cost IoT-Based Soil Monitoring Systems for Precision Agriculture.

• May. 2026: Three papers from the LSACROFT project have been presented at the 49th ICT and Electronics Convention (MIPRO 2026). Watch the presentations here:

  1. • LSACROFT Project Paper Presentations at the 49th ICT and Electronics Convention (MIPRO 2026)

• Three papers from the LSACROFT project have been accepted for presentation at the 49th ICT and Electronics Convention (MIPRO 2026), scheduled to take place from May 25 to 29, in Opatija, Croatia.

  1. A Low-Cost Hardware–Mobile Platform for Comprehensive Soil Monitoring, Crop Phenotype, and Image Dataset Acquisition.

  2. Survey and Design Framework: Comparative Evaluation of Edge Computing Platforms for Low-Cost AI-Enabled Real-Time Soil Monitoring, Dataset Collection, Crop Recommendation, and Fertilizer Optimization in African Settings.

  3. A Comprehensive Survey of Low-Cost IoT-Based Soil Monitoring Systems for Precision Agriculture.

The LSACROFT Project team showcased its Version 1.0 soil data collection device at the CSIR-INSTI Open Day held on 9th April, 2026.

This prototype device is designed to collect key soil parameters in real time, supporting the development of localized datasets for smart agriculture and AI-driven decision-making. The demonstration highlighted how low-cost, field-deployable technology can improve soil monitoring, data collection, and agricultural research.

Event Details:

• CSIR-INSTI Open Day 2026

• Theme: Transforming Innovation Through AI, Precision Technologies, and Open Scientific Knowledge

• Venue: CSIR-INSTI, Accra. 

A link to the device overview:

• LSACROFT Device Overview

The LSACROFT team from CSIR-INSTI completed a field trip to Bolgatanga, where we conducted a device demonstration and training at the experimental farm.

As part of the visit, we recorded a series of videos covering key activities from the trip, including:

• Guided farm tour by Ing. Dr. Raymond Gyaang

• Field experimental setup by Latifa Yakubu

• On-field device demonstration by Elijah Santa

• Workshop on the use of the LSACROFT device by Elijah Santa

• Presentation on the experimental farm and crops by Latifa Yakubu

The seminar focused on Convolutional Neural Networks (CNNs), introducing the fundamental concepts and workflow used in deep learning for image-based analysis.

The session covered the core building blocks of CNNs, explaining how data flows through different stages of the network to extract features and make predictions, with relevance to applications in agriculture and intelligent systems.

The session includes:

1. Introduction to Convolutional Neural Networks and their role in artificial intelligence

2. The convolution operation and feature extraction

3. Activation functions and their role in introducing non-linearity

4. Pooling techniques for dimensionality reduction

5. Fully connected layers and classification

6. End-to-end CNN workflow for image-based applications

The session was presented by Sarah Annang as part of the LSACROFT Project seminar series, with both in-person and virtual participation.

The link to the seminar recording:

• LSACROFT Project Student Lecture: Convolutional Neural Networks (CNNs) and Workflow

As part of the LSACROFT Project, an experimental research site has been established at Mkhanya Research Station in the Eastern Province of Zambia (formerly known as Masumba). The setup and initial experimental activities were supported by the CSA Norhed II Project, with ongoing support from the NSTC–MEST funding to sustain long-term research.

Experimental planting has commenced to support the collection of soil and crop growth data required for the project. The study focuses on a medium maturing maize variety and pigeon pea, which will be monitored throughout their growth cycle to evaluate soil parameters and support dataset development.

The research site will serve as a long-term experimental platform across multiple seasons, supporting both PhD and MSc research on soil sensing technologies. This will enable systematic monitoring of soil conditions and crop performance, contributing to the development of datasets for the LSACROFT project’s work on soil sensing and data-driven agricultural solutions.

The LSACROFT Project team showcased its Version 1.0 soil data collection device at the Ghana Agrotech Fair 2026, held from 17th to 19th March at Black Star Square, Accra.

This prototype device is designed to collect key soil parameters in real time, supporting the development of localized datasets for smart agriculture and AI-driven decision-making. The demonstration highlighted how low-cost, field-deployable technology can improve soil monitoring, data collection, and agricultural research.

Event Details:

• Ghana Agrotech Fair 2026

• Theme: Transforming Agribusiness through Local Innovation and Technology

• Venue: Black Star Square, Accra. 

Link to the demonstration:

• LSACROFT Device V1.0 Demonstration at Ghana Agrotech Fair 2026

As part of the LSACROFT Project, MSc student Latifa Yakubu prepared the experimental farm at Bolgatanga Technical University (BTU). This activity was carried out under the supervision of Ing. Dr. Raymond Gyaang and Prof. Calistus Tengan.

Following land preparation, experimental planting has commenced to collect soil and crop growth data required for the project. The experiment includes selected maize and cowpea varieties that will be monitored throughout their growth cycle to analyze soil parameters and develop a dataset. 

The following crop varieties have been planted for experimentation:

1. Maize varieties

   • Sanzal-sima

   • Wang-daata

   • Bahilifa

2. Cowpea varieties

   • IT-17K-1367-2-3

   • SARI-TUGA

   •  TUZIE-VALLENG

The experimental farm will enable systematic monitoring of soil parameters and crop performance, contributing to the development of datasets required for the LSACROFT project’s research on soil sensing and data-driven agricultural technologies.

The seminar is based on the book TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra‑Low‑Power Microcontrollers by Pete Warden and Daniel Situnayake.

The seminar covers Chapters 1–12 of the book, introducing the fundamental concepts of Tiny Machine Learning (TinyML) and demonstrating how machine learning models can run on small embedded devices such as microcontrollers.

The session includes:

1. Introduction to TinyML and its role in edge AI and embedded systems

2. Fundamentals of machine learning for microcontrollers

3. Data collection and preprocessing for embedded ML applications

4. Building, training, and optimizing lightweight machine learning models

5. Deploying models using TensorFlow Lite on Arduino and ultra-low-power devices

6. Practical applications of TinyML in sensing, IoT, and environmental monitoring

The session was presented by Ralph Tetteh, Elijah Nansuuri, and Sarah Annang and held on 13th March 2026 at CSIR–INSTI, with both in-person and virtual participation.

The link to the seminar recording:

• LSACROFT Project Student Lecture 1: TinyML with TensorFlow Lite on Ultra Low Power Microcontrollers

The first seminar for the year took place at CSIR-INSTI with participants from Bolgatanga Technical University and the University of Zambia joining virtually. This seminar forms part of the capacity-building requirements of the LSACROFT Project, aimed at strengthening skills in data collection, sensor validation, and AI-enabled agricultural research among students and researchers in Ghana and Zambia. The seminar covered:

1. Practical approaches to sensor calibration using the line of best fit

2. Introduction to machine learning–based calibration methods for low-cost sensors

3. A survey and overview of the LSACROFT project theme, objectives, and implementation strategy

Participants included:

• CSIR-INSTI: Dr. Dennis Gookyi, Ing. Dr. Michael Wilson, Dr. Roger Ahiadormey, John Awotwi, Sarah Annang, Yaw Twum, Ralph Tetteh, Elijah Santa, and Michael Dziwornu

• University of Zambia: Prof. Lydia Chabala, Mbita Nakapite, and Chrispin Chansa

• Bolgatanga Technical University: Prof. Callistus Tengan and Ing. Dr. Raymond Gyaang

The link to the seminar recording:

• LSACROFT Project Seminar 1: Sensor Calibration with Machine Learning & Project Overview

The last project meeting for the year took place at the premises of CSIR-INSTI. The agenda of was to brief members on the progress of the hardware device and the mobile software application designed to collect dataset for the project.

Dr. Dennis Gookyi and Prof. Chabala attended the 2025 Global Research Council (GRC) Sub-Saharan Africa Regional Meeting and the Science Granting Councils Initiative (SGCI) Academic Symposium, that took place from 10 to 14 November 2025 in Bulawayo, Zimbabwe. They gave a presentation  on the LSACROFT project.

A project meeting between the Ghana team and the Zambia team took place online to discuss ongoing work. In attendance were Dr. Dennis Gookyi, Prof. Lydia Chabala, Dr. Michael Dziwornu, Ing. Dr. Raymond Gyaang, Dr. Paul Danquah, and Ralph Tetteh.

Two PhD and MSC positions under the LSACROFT project has been announced and shared on various platforms.

The first Seminar meeting for the NSTC and MEST-funded project LSACROFT took place at the premises of CSIR-INSTI. The agenda of the meeting was to build the capacity of students currently undertaking their internship at the institute. During the session, the interns took turns presenting the various research projects they were working on.

Our first project meeting to introduce the project and assign roles to team members took place at CSIR-INSTI. In attendance were Dr. Dennis Gookyi, Ing. Dr. Michael Wilson, Ing. Twum Barimah, Dr. Michael Dziwornu, Mr. John Awotwi, Mr. Emmanuel Kwofie, and Ralph Tetteh.

Dr. Dennis Gookyi and Prof. Lydia Chabala attended the joint online inception meeting organized by NSTC and MEST. They presented the overview of their project titled "Development of a Low-Cost Soil Analysis Device for Real-Time Soil Monitoring, Crop Recommendation, and Fertilizer Optimization (LSACROFT)".