> For the complete documentation index, see [llms.txt](https://sws3009.gitbook.io/sws3009-tutorial/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://sws3009.gitbook.io/sws3009-tutorial/10-july-consultation/group-4.md).

# Group 4

Farming Robot for Plant Inspection and Treatment

#### Objective:

* Develop a farming robot that can autonomously move to specific fields, take photos of plants, inspect for illnesses, and apply the appropriate treatment.

#### Deep Learning:

1. **Plant Inspection**:
   * **Requirement**: Inspect plants one by one, identifying specific illnesses.
   * **Plants and Illnesses**:
     * **Plants**: Three kinds of plants.
     * **Illnesses**: Two specific illnesses for each plant.
   * **Model Development**:
     * Train deep learning models to recognize each type of plant and its associated illnesses.
     * Ensure the model can distinguish between healthy and unhealthy plants.
2. **Treatment Identification**:
   * **Objective**: The model should also determine the appropriate treatment (medication) for identified illnesses.
   * **Implementation**:
     * Develop a system to link identified illnesses to specific treatments.
     * Integrate this information so the robot can inform the robotics side about the necessary action.

#### Robotics:

1. **Autonomous Movement**:
   * **Navigation**: Ensure the robot can autonomously navigate to specific fields.
   * **Localization**: Implement a reliable localization system to accurately position the robot within the field.
2. **Photo Capture**:
   * **Camera System**: Equip the robot with a camera to take high-quality photos of plants.
   * **Data Transmission**: Ensure the robot can send these photos to a central system for deep learning analysis.
3. **Actuation and Treatment Application**:
   * **Actuators**: Equip the robot with actuators capable of performing various actions.
   * **Actions**:
     * **Watering**: Implement a system to water plants as needed.
     * **Nutrient Spraying**: Equip the robot with a mechanism to spray nutrients.
     * **Pesticide Application**: Develop a system to spray pesticides based on identified illnesses.

#### Implementation Steps:

1. **Deep Learning Development**:
   * **Model Training**: Train models to identify plants and specific illnesses.
   * **Treatment Mapping**: Create a database linking illnesses to treatments.
2. **Robotics Development**:
   * **Navigation System**: Develop and test the autonomous navigation system.
   * **Actuation Mechanisms**: Design and implement actuators for watering, nutrient spraying, and pesticide application.
3. **Integration**:
   * **Communication**: Ensure seamless communication between the deep learning system and the robotic control system.
   * **Automation**: Implement a fully automated workflow from plant inspection to treatment application.

#### Next Steps for Students:

1. **Deep Learning Tasks**:
   * Collect and annotate a dataset for the three kinds of plants and their illnesses.
   * Train and validate the deep learning models for accurate plant and illness identification.
   * Develop the treatment recommendation system.
2. **Robotics Tasks**:
   * Design and build the robot with necessary actuators and camera system.
   * Develop the navigation system for autonomous movement.
   * Integrate the actuation mechanisms for applying treatments.
3. **Testing and Iteration**:
   * Test the complete system in a controlled environment.
   * Iterate on the design based on test results to improve accuracy and reliability.

#### Final Considerations:

* **Data Collection**: Ensure a robust dataset for deep learning model training.
* **System Reliability**: Focus on creating a reliable and durable robot that can operate in various field conditions.
* **Scalability**: Design the system to be scalable for different field sizes and plant types.

#### Summary:

* Develop deep learning models to inspect plants and identify specific illnesses.
* Equip the robot with actuators for applying treatments such as watering, nutrient spraying, and pesticide application.
* Ensure seamless integration between the deep learning system and the robotic control system for fully automated plant inspection and treatment.