Nyansapo Artificial Intelligence Challenge




    • Jude Boachie
    • Boniface Delali Dakey

Name of Institution:  Kwame Nkrumah University of Science and Technology


AI-Powered OCR for Document Digitization and Archiving


Our project aims to leverage AI-powered Optical Character Recognition (OCR) technology to revolutionize document digitization and archiving. By automating the conversion of physical documents into digital formats and providing flexible data organization options, we aim to address the challenges faced by libraries, archives, and administrative offices dealing with historical documents and records.

Problem Statement:

Many organizations struggle with the time-consuming and error-prone process of manually digitizing and organizing physical documents. This hinders efficient document management and accessibility, while the lack of customizable data organization options further compounds the problem. Valuable historical documents and records remain underutilized, limiting their potential impact on research, education, and historical preservation. An AI-driven solution is urgently needed to automate the conversion process, offer customizable data organization, and enhance accessibility for efficient document management.

Solution Overview:

Our AI-powered OCR system automates the conversion of physical documents into searchable and editable digital formats. It provides users with the flexibility to organize and arrange data based on specific requirements or predefined patterns. The system utilizes advanced OCR algorithms for accurate document conversion and incorporates AI algorithms for pattern recognition and metadata extraction. Advanced search functionality enables efficient retrieval of information. The solution is scalable and integrates with existing systems for seamless digital document management.

Impact and Benefits:

Implementing our solution in Ghana will have significant impact and benefits. It will enable libraries, archives, and administrative offices to convert physical documents into digital formats, enhancing the accessibility and preservation of historical records. This promotes research, education, and knowledge dissemination, contributing to the country’s cultural heritage and historic preservation efforts. Customizable data organization and advanced search capabilities improve document retrieval, saving time and effort for users. Efficient document management enhances operational efficiency and productivity, empowering institutions to serve their communities better and support informed decision-making.


Our AI-powered OCR system has the potential to transform document digitization and archiving, benefiting organizations in Ghana dealing with physical documents. By automating the conversion process, improving accessibility, and streamlining information retrieval, our solution supports the preservation of historical records, enhances efficiency, and facilitates knowledge dissemination. Implementing this project will contribute to preserving Ghana’s cultural heritage, improving accessibility, and ensuring efficient management of historical information.



    • Joseph B. Tandoh
    • Farouk Dayshini
    • Michael S Hayford

Name of Institution:  Ghana Space Science Institute


Title: Developing a Predictive Model for Water Quality for Ankobra River using Machine Learning Techniques


Water quality degradation in Ghana’s Ankobra River is a significant concern linked to pollution from mining, agriculture, and industrialization. Preserving the river’s water quality is crucial due to its importance for irrigation, domestic use, and fishing. Developing a machine learning-based predictive model can enhance understanding of degradation factors and enable targeted interventions. Studies by Owusu-Boateng et al. (2018), Ofori et al. (2019), and Abban et al. (2020) confirm the adverse impact of human activities on Ankobra River’s water quality. Effective regulations are needed to reduce pollution from mining, agriculture, and domestic sources and maintain acceptable water quality levels. The predictive model empowers stakeholders with valuable insights to implement interventions and protect the river’s ecosystem.

Research Question:
Can a predictive model for water quality at Ankobra River be developed using machine learning techniques, and how accurately can it forecast future water quality?


  1. Sample and analyze water quality data from the Ankobra River over the last five years (from literature and onsite).
  2. Develop a predictive model and evaluate its accuracy using machine learning techniques to forecast future water quality.
  3. To identify the primary factors contributing to water quality degradation and recommend interventions to improve water quality.

This research will use a quantitative approach to collect and analyze water quality data from the Ankobra River. Water quality parameters such as pH, dissolved oxygen, biochemical oxygen demand, chemical oxygen demand, temperature, and total dissolved solids will be measured and recorded at various points along the river. The data will be collected over a period of five years. The data collected will be used to develop a predictive model using machine learning techniques such as linear regression, decision trees, and artificial neural networks. In addition to machine learning, remote sensing techniques will play a vital role in this research. Satellite imagery will provide valuable information on water quality-related features. The model will be trained using the historical water quality data, and its performance will be evaluated using a hold-out validation approach. The accuracy of the model will be evaluated by comparing the forecasted water quality with actual water quality data collected during the validation period. By integrating remote sensing data with machine learning algorithms, a comprehensive understanding of the spatial and temporal dynamics of water quality degradation can be achieved.

Expected outcome:
The expected outcome of this research is a predictive model that can forecast future water quality at Ankobra River using machine learning techniques.

The development of a predictive model for water quality at Ankobra River using machine learning techniques would provide valuable insights into the factors contributing to water quality degradation and enable targeted interventions to improve water quality.



    • Gare Lawson
    • Razak Moazu
    • Daniel Ewusi-Essel

Name of Institution:  Accra Technical University


Title: Energy-Efficient Building Management System Powered by Artificial Intelligence (A case study of the Hostel Facilities of Accra Technical University)


Energy efficiency has become a crucial factor in building design and management, considering the negative impact of greenhouse gas emissions on the environment. With the increasing demand for energy, it is necessary to find innovative ways to manage and optimize energy consumption in buildings and hostels of Accra Technical University. The integration of artificial intelligence (AI) technology in building management systems can aid in reducing energy consumption and achieving energy efficiency for the Accra Technical University hostels.


The objective of this concept note is to propose an energy-efficient building management system that utilizes AI technology to measure power usage and predict energy consumption in the hostel facilities of Accra Technical University.


The proposed system will use AI algorithms to analyze data collected from sensors placed in different parts of the building of the hostels. The data collected will be used to predict the amount of energy required to maintain the building’s temperature, lighting, and ventilation at optimal levels. The system will also monitor the energy consumption of different devices and appliances within the building, such as air conditioning systems, lighting, and electrical equipment.


The proposed system will have several benefits, including:

Energy Savings: The system will help to reduce energy consumption by optimizing the use of energy in the building.

Cost Savings: The reduction in energy consumption will lead to lower utility bills and operational costs.

Improved Comfort: The system will help to maintain optimal temperature and lighting conditions, providing a comfortable environment for building occupants.

Environmental Sustainability: The reduction in energy consumption will contribute to environmental sustainability by reducing greenhouse gas emissions.


An energy-efficient building management system powered by AI technology has the potential to revolutionize the building management industry by optimizing energy consumption, reducing costs, and improving environmental sustainability. With the increasing demand for energy, it is necessary to embrace innovative solutions that promote energy efficiency and sustainability.



    • Jessica Oparebea Appiah
    • Ebenezer Domey Appiah
    • Elvis Ampoh

Name of Institution:  University of Ghana




I am writing to submit a concept proposal for the Nyansapo AI Challenge 2023, the project which aims to address the critical issue of detecting and treating the top three diseases affecting cocoa and maize crops in Ghana.


Ghana is a major producer of cocoa and maize, two crops that are essential to the country’s economy. However, cocoa and maize crops are susceptible to a number of diseases, which can cause significant yield losses. In Ghana, the three most common cocoa diseases are black pod, swollen shoot, and witches’ broom. The three most common maize diseases are maize streak virus, maize rust, and northern leaf blight.

Detecting and diagnosing these diseases poses challenges, particularly for small-scale farmers who may lack the necessary expertise. This often results in delayed treatment, leading to decreased crop yields. Moreover, the financial burden of treating cocoa and maize diseases can be overwhelming for small-scale farmers.


The proposed mini-project aims to develop an AI model capable of accurately detecting and classifying the three most common diseases affecting cocoa and maize crops.

This model will be integrated into a user-friendly mobile application targeted toward small-scale farmers in Ghana. By leveraging deep learning algorithms, the model will be trained on a comprehensive dataset comprising images of infected cocoa and maize plants, allowing it to learn and recognize the distinct features associated with each disease.

The mobile app will be designed with simplicity in mind, ensuring ease of use for farmers. The model will help identify the specific disease and provide an appropriate treatment plan.


The project is intended to function in the following ways.

Detect and diagnose diseases early, which will help to reduce yield losses.

Get treatment advice for diseases, which will help to improve yields.

Save money on the cost of treating diseases.

The app will also help to improve the efficiency of agricultural extension services in Ghana. Extension agents will be able to use the app to provide farmers with information about diseases and treatment plans.

The project has the potential to transform disease management in cocoa and maize farming in Ghana. By integrating AI technologies into the agricultural sector, we can bring about meaningful change and improve the livelihoods of farmers.

The team is excited to further develop and implement this project, and we look forward to the opportunity to showcase our innovative solution.



    • Bawa Daniel Dangema
    • Hickman Mohammed
    • Adamptey Ebenezer

Name of Institution:   University of Energy and Natural Resources, UENR





Pipelines are the most efficient and safest means for the transportation of oil, gas, and refined petroleum products. Potentially severe consequences of pipeline failures make reliability and risk assessment an essential aspect of safe operation. Corrosion is one of the major causes of pipeline failure, seriously affecting pipelines operation, according to the 11th Report of the European Gas Pipeline Incident Data Group ((EGIG), 2020), 26.63% of gas pipeline incident were caused by corrosion.

Problem Statement

  • Current pipeline inspection practices rely heavily on manual inspection, which is dangerous and unreliable. This poses risks to human safety and the environment as undetected anomalies can result in disastrous consequences.
  • A more reliable solution is needed to ensure efficient and cost-effective inspections so that anomalies can be detected and resolved in a more timely and effective manner.

Proposed Solution

A proposed system aims to detect anomalies such as corrosion and cracks in oil and gas pipelines using a combination of hardware and software.  The system will use robotic navigation techniques to move inside the pipeline and capture images of pipeline surfaces. Mapping of the pipeline’s path will be done with an Initial Measurement Unit (IMU) and Visual simultaneous localization and mapping (vSLAM). Real-time data will be transmitted using LoRaWAN technology. Machine learning, specifically supervised learning, will be used to analyze captured images and predict the presence and severity of corrosion and cracks in the pipeline.

Impact of the solution

Using pipeline inspection robot (vehicle) to capture images of corroded and cracked internal surfaces of oil and gas pipelines and training a supervise learning model to predict the corrosion level of the pipelines will go a long way to help oil and gas industries to detect the deterioration level of their pipelines to avoid any unforeseen circumstances that could lead to disaster and loss of lives as well as posing harmful effects on the environment. This project will be a revolution in the oil and gas sector contributing to the UN sustainable development goals i.e., Goal 8 (Decent work and economic growth), Goal 9 (Industry, Innovation and Infrastructure), Goal 14 (Life below water) and Goal 15 (Life on Land).



    • Nyamekesse Samuel
    • ‪‪‪‪‪‪‪‪‪‪‪‪Charles Roland Haruna (Supervisor)
    • Peter Baah Kpabitey

Name of Institution:  University of Cape Coast



AI-Based Healthcare Application for Improved Access and Emergency Response

Problem Statement:

Limited access to fast medical care, reliable diagnosis, and quick emergency responses to patients is a critical issue in healthcare systems. This problem particularly affects individuals in remote areas and underserved communities, resulting in delayed treatment and compromised health outcomes.

Proposed Solution:

We propose developing an AI-based healthcare application that aims to address the problem of limited access and emergency response in healthcare. The application will leverage AI algorithms and real-time data analysis to provide fast medical care services, accurate diagnosis, and efficient emergency responses to patients in need.

Key Features and Functionalities:

  1. AI-Powered Diagnosis:

Utilizing machine learning algorithms, the application will provide an accurate and timely diagnosis by analyzing patient symptoms and medical history, enabling personalized treatment recommendations.

  1. AI-Powered Personalized Health Insights:

Through data analytics, the application will provide patients with personalized health insights, preventive care recommendations, and reminders for medication adherence, promoting overall wellness.

  1. Fast Medical Care Services:

The application will connect patients with licensed medical professionals, allowing them to schedule virtual consultations and receive prompt medical advice and prescriptions both from the AI-integrated model and health professionals.

  1. Real-Time Emergency Response:

The application will incorporate a GPS or map locator to swiftly identify the patient’s location in times of emergency, facilitating quick medical assistance and ambulance dispatch.

Expected Outcomes:

By implementing this AI-based healthcare application, we anticipate the following outcomes:

  1. Improved access to medical care for individuals in remote areas and underserved communities.
  2. Timely and accurate diagnosis, leading to appropriate and prompt treatment.
  3. Enhanced emergency response, ensuring quick medical assistance during critical situations.
  4. Increased patient satisfaction, improved health outcomes, and reduced healthcare disparities.



    • Prince Kyeremanteng Samuel
    • Dr. Kofi Sarpong Adu-Manu
    • Prince Ofori

Name of Institution:  University of Cape Coast



EnergiaAI: Transforming Ghana’s Grid Monitoring with Artificial Intelligence


Reliable electricity is vital for economic development and citizens’ well-being. However, Ghana faces persistent power shortages, causing frequent disruptions that harm businesses and burden the national economy. To overcome this challenge, we employ real-time monitoring of the power grid using Artificial Intelligence to detect potential failures before they happen so that we can effectively combat them.


The power grid in Ghana is susceptible to the challenges of instability, damage of equipment, and low voltage that can result in power outages, which have negative impacts on economic development and the quality of life of citizens. The country utilizes a combination of manual inspections such as data collection from substations, and Supervisory Control and Data Acquisition (SCADA) systems, which can be time-consuming and costly.


The proposed solution, EnergiaAI, leverages the power of artificial intelligence (AI) to actively monitor Ghana’s power grid, proactively identifying potential failures before they result in disruptive power outages. By harnessing data from diverse sources such as power generation, transmission, distribution, weather, equipment, maintenance, and outages, the system employs advanced machine learning algorithms to meticulously analyze this data. Through this AI-driven approach, EnergiaAI detects patterns and anomalies that could signify impending failures or malfunctions within the power grid, thus playing a pivotal role in safeguarding its stability and preventing service disruptions.


EnergiaAI is expected to yield the following results:

  • Reduce the frequency and duration of power outages by foretelling when an outage is to occur.
  • Enable early prediction of equipment damages in the power grid to minimize maintenance costs and prevent catastrophic failures.
  • Enable real-time alerts and notifications, empowering power grid operators and maintenance personnel to promptly address issues and ensure seamless operations.


The implementation of EnergiaAI will involve the following steps:

  • Data Collection:

At this stage, we will collect and aggregate data from various sources, including power generation, transmission, distribution, weather, equipment, maintenance, and outage data.

  • Data Processing:

Here, we will clean, filter, and process the data to prepare it for analysis.

  • AI model development:

Developing the AI model to analyze the data and to identify potential failures or malfunctions in the power grid.

  • Integration with existing systems:

 Integrate the AI model into any existing power grid monitoring system.