Guzel Mustafa, Turan Bulent, Kadioglu İzzet, Sin Bahadir, Basturk Alper, Khaled R. Ahmed

How to Affect the Number of Images on the Success Rate for Detection of Weeds with Deep Learning

The detection of weeds with computer vision without the help of an expert is important for scientific studies and other purposes. The images used for the detection of weeds are recorded under controlled conditions and used in image processing-deep learning methods. In this study, the images of 3-4-leaf (true-leaf) periods of the wild mustard (Sinapis arvensis) plant, which is the critical process for chemical control, were recorded from its natural environment by a drone. The datasets were included 50-100-250-500 and 1 000 raw images and were augmented by image preprocessing methods. Totally 12 different augmentation methods used and datasets were examined for understand how to affects the numbers of images on training-validation performance. YOLOv5 was used as a deep learning method and results of the datasets were evaluated with the Confusion Matrix, Metrics-Precision, and Train-Object Loss. For results of Confusion Matrix where 1 000 images gave the highest results with TP (True Positive) 80% and FP (False Positive) 20%. The TP-FP ratios of 500, 250, 100 and 50 image numbers were respectively; 65%-35%, 43%-57%, 0%-100% and 0%-100%. With 100 and 50 images, the system did not show any TP success. The highest metrics- precision ratio was found 92.52% for 1 000 images set and for 500 and 250 image sets respectively; 88.34% and 79.87%. The 100 and 50 images datasets did not show any metrics-precision ratio. The minimum object loss ratio was 5% at 50th epochs in the 100 images dataset. This dataset was followed by other 50, 250, 500, and 1 000 images respectively; 5.4%, 6.14%, 6.16%, and 8.07%.

PDF

___

Arnason THRJaJT. 1981. Use of plants for food and medicine by Native Peoples of eastern Canada. Can. J. Bot., Issue 59: 2189-2325.
Bàrberi P. 2002. Weed management in organic agriculture: Are we addressing the right issues? Weed Res., Issue 42: 177-193.
Başaran BKYKİKDTHAM. 2016. The Effect of 2,4 D Acid Dimethylamin Against Broadleaf Weeds Applied at Different Phenological Periods on Grain Yield and Some Yield Components of Common Wheat (Triticum aestivum L.). Turk J Weed Sci, Issue 19(2): 1-9.
Blackshaw REAGWaDJ. 1987. Interference of Sinapis arvensis L. and Chenopodium album L. in spring rapeseed (Brassica napus L.). Weed Res., Issue 27: 207-213.
Güzel M, Şin B, Turan B, Kadıoğlu I. 2021. Real-Time Detection of Wild Mustard (Sinapis arvensis L.) With Deep Learning (YOLO-v3). Fresenius Environmental Bulletin, Issue 30 11/A20201, pp. 12197-12203.
Hasan A. et al. 2021. A survey of deep learning techniques for weed detection from images. Comput. Electron. Agric., Issue 184.
Kamilaris A, Prenafeta-Boldu F. 2018. Deep learning in agriculture: A survey. Comput. Electron. Agric., Issue 147: 70-90.
Koirala A, Walsh K, Wang Z, McCarthy C. 2019. Deep learning Method overview and review of use for fruit detection and yield estimation. Comput. Electron. Agric., Issue 162: 219- 234.
Liakos K. et al. 2018. Machine Learning in Agriculture: A Review. Sensors, Issue 18: 2674.
Mavridou E. et al. 2019. Machine Vision Systems in Precision Agriculture for Crop Farming. J. Imaging, Issue 5:89.
McMullan PMDJKaDDR. 1994. Effect of wild mustard (Brassica kaber) competition on yield and quality of triazine-tolerant and triazine-susceptible canola (Brassica napus and Brassica rapa). Can. J. Plant Sci., Issue 74: 369-374.
Mulligan GAaBLG. 1975. The biology of Canadian weeds: Sinapis arvensis L. Can. J. Plant Sci, Issue 55: 171-183.
Nepal U, Eslamiat H. 2022. Comparing YOLOv3, YOLOv4 and YOLOv5 for Autonomous Landing Spot Detection in Faulty UAVs. Sensors, Issue 22: p. 464.
Rollins RC. 1981. Weeds of the Cruciferae (Brassicaceae) in North America. J. Arnold Arbor, Issue 62: 517-540.
Sabzi S, Abbaspour-Gilandeh Y, Arribas J. 2020. An automatic visible-range video weed detection, segmentation and classification. Heliyon, Issue 6.
Şin B. 2021. The determination of tribenuron methyl resistance of wild mustard (Sinapis arvensis L.) collected from wheat fields located in Amasya, Çorum, Tokat and Yozgat province. Tokat: Tokat Gaziosmanpaşa University, Institute of Science and Technology, Ph.D. Thesis.
Şin B, Kadıoğlu İ. 2021. Trp-574-Leu mutation in wild mustard (Sinapis arvensis L.) as a result of als inhibiting herbicide applications. PeerJ, Issue 9: e.
Su W. 2020. Advanced Machine Learning in Point Spectroscopy, RGB- and Hyperspectral-Imaging for Automatic Discriminations of Crops and Weeds: A Review. Smart Cities, Issue 3: 767-792.
Thomas AGaWRF. 1984. Weed survey of Manitoba cereal and oilseed crops 1978, 1979 and 1981. Weed Survey Series Publ., Issue 84-1: 230.
Tian H. et al. 2020. Computer vision technology in agricultural automation—A review. Inf. Process. Agric., Issue 7: 1-19.
Uygur FKWWH. 1986. Definition of Important Weeds in Wheat- Cotton Planting System of Çukurova Region. PLITS, Issue 1986/4(1): 169.
Weng Y. et al. 2019. A survey on deep-learning-based plant phenotype research in agriculture. Scientia Sinica Vitae, Issue 49: 698-716.
Wu Z, Chen Y, Zhao BKX, Ding Y. 2021. Review of Weed Detection Methods Based on Computer Vision. Sensors, Issue 21.
Yuan H, Zhao N, Cheng M. 2020. Review of Weeds Recognition Based on Image Processing. Trans. Chin. Soc. Agric. Mach., Issue 51: 323-334.
Zhang S, Huang W, Wang Z. 2021. Combing modified Grabcut, K-means clustering and sparse representation classification for weed. Neurocomputing, Issue 452: 665-674.