Identification and counting of brown planthopper in the paddy field using image processing techniques / Nur Atiqah Nasser Shah
Massive paddy crops is lost every year, due to weather condition, plant disease and pest attack. Pest attack is among the main threat that reducing the quality and quantity of the paddy crops. Pesticide is used to control and eliminate the pest in paddy field. In order to determine the quantity of p...
Saved in:
| Main Author: | |
|---|---|
| Format: | Student Project |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://ir.uitm.edu.my/id/eprint/38094/1/38094.pdf http://ir.uitm.edu.my/id/eprint/38094/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Massive paddy crops is lost every year, due to weather condition, plant disease and pest attack. Pest attack is among the main threat that reducing the quality and quantity of the paddy crops. Pesticide is used to control and eliminate the pest in paddy field. In order to determine the quantity of pesticide used, traditional method is based on manual counting where sticky trap is used to trap the pest and after that the pest was counted manually. However this method is tedious and time consuming due to many crops generally found in paddy field. This can lead to imprecise counting through the process and delay in order to gain the accurate count. This paper proposed an image processing technique and artificial intelligent as an alternate method for identification and counting of pests. The suggested pest detection consist of four (4) consecutive steps; image acquisition, image segmentation, feature extraction and detection. The detection step contains two processes that are classification and counting of pest. A k-nearest neighbour (kNN) classifier is used in this step. The process is developed and implemented using MATLAB. Simulation results using 10 pest images indicated that the classification performance of brown planthopper using kNN provides better performance compared to decision tree. The kNN achieved precision 0.97, recall 0.96, accuracy 0.97 and F score 0.96. The system will be used in assisting paddy field worker to predict the type and amount of pesticide to be used for the pest control. |
|---|