Article Preview
TopIntroduction
The agriculture field contains so many challenging applications for yield prediction. In Chhattisgarh, farmers mostly depend upon rice production for their livelihoods and it is advantageous over the Nation also. Therefore, production growth should be proper even with some uncertainties like weather disturbance, pests’ attack, virus attacks, bacterial and fungal attacks etc. The selection of a method for profitable production growth requires plant phenotyping traits analysis. Recently, applied various popular phenotypic traits also have significant limitations for cost, performance, space and time coverage (Li et al., 2020). One of the other limitations for learning patterns from the plant spectra is a crucial task for predictive plant phenotyping applications discussed by Rehman et al. (2020). Some other challenging applications using machine learning approaches has listed as: digitized regularization, smart farming using IOT, irrigation method, precision agriculture using Artificial Intelligence, and market prediction analysis discussed by Sharaff and Choudhary (2018) etc. These traits having major utility for the reduction of the complex genome executions over real time data analysis in the field of agriculture.
The main contribution of the proposed work includes:
- 1.
Early disease detection for high yield production.
- 2.
Encouragement towards the interdisciplinary approach for smart farming by using computer vision techniques.
- 3.
Field farming needs digitization/computer-aided techniques, so that performance rate gets increases. It avoids the complexity of manual calculation, time management, hardware management, and so on.
- 4.
The proposed work improves the phenotyping traits analysis over complexity of genomic selection. Traits analysis (Tong and Nikoloski, 2021) of plant has multiple tasks, but it is exclusively an abstract structure over the genotyping to produce fruitful results by applying it.
- 5.
It gives the vision towards the heterogeneous environment. Prediction over the environmental uncertainties like rainfall prediction, temperature management (field conditions are heterogeneous by nature that will not give identical results), weather forecasting and avoiding global warming problem over the world.
- 6.
Management of Eco-system is having all the above dependencies over smart farming or forest-scale phenotyping (Bombrun et al., 2020) etc. Smart farming requires the computer vision techniques to limit the error rate calculation and maximize the early prediction rate.
- 7.
Utility of machine learning algorithms make a trend for working efficiently in the field of agriculture.
Figure 1 represents the significance of plant phenotyping over genotyping data analysis. Genotyping creates a complexity to analyze the plant features for regular analysis. In genome data analysis, GWAS relates the gene functionality of plant data. In contrast, plant phenotyping data analysis describes with some of the plant feature analysis such as height, leaf disease, biomass, LAI (leaf area index), CCC (canopy chlorophyll content), CT (computed tomography) (Liu et al., 2017; Zhang et al., 2018), color index based segmentation and so on. Plant phenotyping traits analysis contains identification or prediction of crop diseases with multi-domain image processing task.
Figure 1.
Plant Phenotyping Data Analysis over Genome Data
In this paper, every descended section is distributed for their respective purposes. According to the second module, literature survey has been discussed, which gives knowledge about the latest concerning area and problem statement for the future work. The third module describes the methodology and process of the proposed work with the enhanced result analysis. The fourth module has an experimental setup to process step-wise execution of all the learning phases. The next section relates tabular representation of the result analysis and the last section concluded this proposed work, which indicates utility towards yield prediction by applying machine learning techniques over the real time dataset.