Real time pest detection and fumigation using machine learning
Pakistan's growth and its economy greatly depend upon its export through the agriculture sector. The main problem faced in the agricultural sector is the deterioration of crop quality as pests destroy the crop quality. To counter this problem, we aimed to design a robotic arm that is capable of dete
2025-06-28 16:28:55 - Adil Khan
Real time pest detection and fumigation using machine learning
Project Area of Specialization Artificial IntelligenceProject SummaryPakistan's growth and its economy greatly depend upon its export through the agriculture sector. The main problem faced in the agricultural sector is the deterioration of crop quality as pests destroy the crop quality. To counter this problem, we aimed to design a robotic arm that is capable of detecting pests present on the plant and spraying pesticides on that specific part of the plant. It saves the crop quality and maintains its health. This robotic arm is capable of scanning multiple plants present in the field by scanning the whole plant from top to bottom to detect and kill any pest present.
Project ObjectivesThe following are the aims and objectives of our project:
1. Secure plant health: The quality of the plant is the major factor, without that it is not possible to meet the food demands and it also directly impacts the export of our country. So we aimed to design a robot arm that helps in protecting the plants from pests.
2. Designing a device for pest detection: The aim is to design a device that detects and sprays pesticides if the presence of a pest is detected on the plant and maintains the health and growth of the plant.
3. Reduction of human workload: Spraying pesticides on plants is a tedious task and it is very harmful to human health. Our designed robot will perform the detection task as well as kill the pest to reduce human work.
4. Saving of time: As compared to manual spraying, machines will perform the task in a lesser amount of time with more reliability and accuracy.
5. Limited usage of pesticide spray: The designed robot is capable of spraying the pesticides to the particular area where pests are detected. This will reduce the wastage of pesticides and cost.
The project implementation consists of four main parts:
1. Artificial intelligence ( pest detection): Here we have applied the concept of machine learning known as transfer learning using Jetson nano kit. We have trained Jetson nano using a data set of multiple pests of a tomato plant, by which it is able to detect pests on the plant.
2. Pest killing system: In this part, the trained model communicates with the Arduino to spray pesticides on the area in which it has detected the pests. The sprayer consists of a DC pump which is activated when a bounding box is created over the pest and sprays the pesticide on that same location, where the pest is detected
3. Robotic-arm controlling: This part consists of a robotic arm that moves vertically and rotates about its axis so that it can scan the entire plant from top to bottom. This mechanism is controlled by Arduino through DC motors.
4. Four-wheeler moving robot: This part consists of a four-wheeler base on which the whole structure of the robotic arm is implemented. It consists of four DC gear motors which are controlled by Arduino, that help the robotic arm to move around the plant so that the whole plant is scanned
Benefits of the ProjectHere is the list of benefits of the project:
1. Maintaining crop quality: This device ensures crop quality by detecting and killing pests detected on the plant as it scans the plant from top to bottom and hence maintaining both plant and crop quality.
2. Reducing human effort: This device can scan the plant properly which is difficult to do manually, which saves the human effort.
3. Saving time: This device can perform detection and killing parts in much lesser time as compared to humans checking the whole plant manually.
4. Limited amount of pesticides is used: This device saves the quantity of the pesticide spray as it only sprays on the pest present in the specific area of the plant. This saves the quantity of spray as with manual spraying the pesticides are sprayed all over the plant.
Technical Details of Final DeliverableHere are the details of the final deliverable:
This project consists of two main parts of hardware and software.
Software:
- Artificial Intelligence ( pest detection)
- Python programming for tracing and killing pests
- Arduino coding for controlling robotic arm and four-wheeler chassis.
- IoT for monitoring plant health
Hardware:
- Logitech camera for pest detection
- DC pump for spraying
- Vertical sliding robotic arm with on-axis rotation
- Four wheeler chassis for moving the robot in the field of plants
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 65070 | |||
| Jetson nano | Equipment | 1 | 22000 | 22000 |
| Raspberry pi camera | Equipment | 1 | 6000 | 6000 |
| Logitech webcam | Equipment | 1 | 3300 | 3300 |
| Chassis and assembly | Equipment | 1 | 5230 | 5230 |
| Sd cards | Equipment | 1 | 3250 | 3250 |
| Power supply and cables | Equipment | 1 | 1600 | 1600 |
| T8 rod and nut | Equipment | 1 | 2050 | 2050 |
| Glue and adhesive | Equipment | 2 | 265 | 530 |
| Stepper drivers with gearbox | Equipment | 3 | 810 | 2430 |
| Servo motors | Equipment | 2 | 450 | 900 |
| Wheels and motors | Equipment | 1 | 3300 | 3300 |
| Touchscreen lcd | Equipment | 1 | 9000 | 9000 |
| Dc motors | Equipment | 1 | 1300 | 1300 |
| Artificial pests | Equipment | 1 | 1150 | 1150 |
| Stepper motor | Miscellaneous | 3 | 330 | 990 |
| Belts and gears | Equipment | 1 | 1040 | 1040 |
| Arctificial leaves | Miscellaneous | 1 | 500 | 500 |
| Sliding channel | Miscellaneous | 1 | 500 | 500 |