Phenological Property Consideration for Crop and Weed Discrimination Technologies: A Review

Author: Jyoti Lahre

Download PDF

Abstract

Unwanted plants grow non-uniformly, autonomously in the field, and compete with the major crop called weeds. It competes with the main crop for sunlight, nutrients, water, and space and grows faster. These effects affect the growth rate of crop seedlings, eventually resulting in crop yield reduction. Weed control is very critical to crop production. Several studies have examined the yield loss associated with weed competition. Due to the phenotypic similarity between some crops and weeds as well as changing weather conditions, it is challenging to identify and design an automated system for general weed detection. Many research studies documented various weed discrimination, identification, and control techniques. These recognition mechanisms could be mechanical or physical for intra-row weeding. Image segmentation, height/stalk identification, machine vision systems, sensor-based approaches, RTK-GPS based systems, etc. It is better to control weeds effectively. These advancing technologies promise agriculture improvement with fewer labor-intensive tasks. The more challenging area of intra-row mechanical weeding with manually operated weed control is labor-intensive and time-taking. Along with various discrimination solutions for weed control discovered in industry and the research community, the state of the art in automated mechanical weeding is being explored. An automated technique includes data acquisition and processing. Data processing includes typical plants’ morphological trait extraction and estimation based on a multi-level region segmentation method. Automatic morphological traits are compared with manually measured values. The proposed method's robustness and low time cost for different plants, show potential applications for real-time plant measurement and high-throughput plant phenotyping. In this paper, we study different methods or techniques for weed recognition.

Keywords

Crop production, recognition mechanism, RTK-GPS, morphological trait, segmentation, plant phenotyping

Conclusion

Weed identification and removing is major challenge for intra row crop field. Most of weeds have same characteristics of main crop plant, which is major problem for site specific weed management. In farmer’s point of view, the reduction of herbicide uses by different intercultural practices, and investment in relative expensive and complex equipment, without an expectation of increased yield, there should be an acceptable technology used. The main benefits are the savings in production means (herbicide costs) and improved autonomy. Therefore, the introduction of new systems needs to be properly supported and maintained in order to successfully introduce them to farmers. The image segmentation, color and shape identification, active shape models and UAV imagery are satisfactory at their work. Height and stalk location are complex but precise at results. Ground-level sensors offer very high spatial resolution, and therefore the potential ability to apply classification to classes comprising only one plant species It appears possible that small innovative companies may be the primary source of new weed management technology in the future. Based on the vast improvements in robotics and processing, it would appear that the future of automation in weed control is very promising. Given the high-level performance in this paper, it was demonstrated that the reviewed methods are suitable for the ground-based weed identification in vegetable plantation under various conditions, including varied illumination, complex backgrounds as well as various growth stages and has application value for the sustainable development of the vegetable industry. Future work will be conducted to identify weeds in in-situ videos. Meanwhile, it would also be interesting to evaluate the accuracy reached in the detection of vegetables by optimizing the deep learning model.

References

-

How to cite this article

Jyoti Lahre (2023). Phenological Property Consideration for Crop and Weed Discrimination Technologies: A Review. Biological Forum – An International Journal, 15(3): 450-465.