Feature Extraction for the Discrimination of Crop and Weed in Digital Images using Open CV and Python
Senthilkumar K.1, Sathiyaraj Subramaniyam2, Sharmila S.3, Illakiah H. R.4
1Senthilkumar K*., Department of Mechatronics Engg, Coimbatore Institute of Engg and Technology, Coimbatore, Tamilnadu, India.
2Sathiyaraj Subramaniyam, Department of Crop Physiology , Adhiyaman College of Agriculture and Research, Tamilnadu, India.
3Sharmila S., Department of Mechatronics Engg, Coimbatore Institute of Engg and Technology, Coimbatore, Tamilnadu, India.
4Illakiah HR , Department of Mechatronics Engg, Coimbatore Institute of Engg and Technology, Coimbatore, Tamilnadu, India.
Manuscript received on January 23, 2020. | Revised Manuscript received on February 05, 2020. | Manuscript published on February 30, 2020. | PP: 3461-3465 | Volume-9 Issue-3, February 2020. | Retrieval Number: B3245129219/2020©BEIESP | DOI: 10.35940/ijeat.B3245.029320
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Variable rate herbicide spraying technology has become integral part of precision agriculture and this system works based on the weed density map of agriculture field. To improve the accuracy of crop/weed discrimination process this paper presents different image processing techniques. Edge detection process for obtaining contour is performed by using sobel operator with 5X5 gradient operator and canny edge detector. Grayscale morphological operations are performed to remove gray overlap due to background of the image in order to improve the accuracy of the segmentation process. In order to check the discrimination accuracy and extracting image features, the experiment was performed on 100 images of maize plant and weed plant leaves. From the experimental results, it is concluded that the proposed method can accurately extract leaf parameters for discrimination process with soil background.
Keywords: Edge detection, Gray scale morphology, crop/weed discrimination, Image features, Image classification