Disease Detection in Coffee Plants using Computer Vision

Local Chapter Ethiopia Local Chapter

Coordinated byEthiopia ,

Firi Berhane
, Musse Woldesemayat

Status: Completed

Project Duration: 28 Jan 2023 - 11 Mar 2023

Open Source resources available from this project

Github Repository

Project background.

Ethiopia is the birthplace of coffee and the current source of the best coffee in the world. The most diverse area of Ethiopian agriculture is found in the coffee-growing sector. Ethiopia produces the most coffee in Sub-Saharan Africa and ranks fifth globally, producing between 7-10% of the world’s total coffee production.

Coffee is grown and consumed on every continent in the world. With a contribution of 25% to 30% of all foreign exchange earnings, coffee is Ethiopia’s top export product. At least 15 million people rely on coffee for their livelihood. Over 80% of coffee growers are small-scale farmers. Whole towns depend on coffee for livelihood, and small producers struggle to make a living while dealing with growing environmental challenges.

The problem.

Information obtained from the Ethiopian Coffee and Tea Authority revealed that coffee berry disease (CBD), CWD and coffee leaf rust (CLR) are the three major fungal diseases of Arabica coffee, reducing coffee production and consumption in the country. The approach employed for illness surveillance is observation with the naked eye, which is time-consuming, expensive, and requires significant competence. Therefore, it is important to automatically identify the diseases without the need for experts.

We can leverage the use of deep learning, object detection, and image classification to solve this problem.

Deep learning methods have been introduced for the detection of different types of coffee plant diseases caused by pests and pathogens. These diseases can be classified by machine learning techniques like segmentation, and classification along with the estimation of the severity of stress.

Project goals.

1. Study the types of diseases that commonly exist. 2. Source for train datasets. 3. Carryout data preprocessing 4. Develop a Deep Learning model. 5. Carry out inference with the trained model using test data. 6. Develop mobile app.

Project plan.

  • Week 1

    Data Collection and study the types of diseases that commonly exist

  • Week 3

    Building the model

  • Week 4

    Training the model

  • Week 5

    Training the model and Testing model with test data

  • Week 6

    Develop a mobile app

Learning outcomes.

– Collection of Data.
– Data preprocessing.
– Build a Deep Learning model.
– Mobile Development.
– Transfer Learning.
– Deep CNN.

Share project on: