Description:Field of the Invention
The AI-Driven Plant Health Monitoring and Crop Yield Forecasting Using YOLO and Random Forest falls under the field of Agricultural Technology (AgriTech) and Artificial Intelligence (AI) Applications in Agriculture. It combines advancements in Machine Learning (ML), Deep Learning (DL), and Computer Vision to address challenges in precision agriculture, crop yield forecasting and plant health monitoring. This invention aims to revolutionize agricultural practices by integrating data-driven decision-making tools for sustainable farming and improved productivity.
Background of the Invention
Agriculture is one of the most important parts of human civilisation and contributes to the global economy. This sector faces challenges like unpredictable weather conditions, presence of plant diseases, and inability to use resources properly. Such conditions often lead to reduced yield, monetary loss and also causes harm to the environment.
Traditional agriculture methods rely on manually observing the weather and historical trends. But such data is often insufficient to address the issues of modern agriculture. We are going to use new technologies like machine learning models and deep learning models to enable data driven agriculture. There are models that exist that perform crop yield prediction and leaf disease detection separately but we aim to integrate them into one to get more precise results.
The innovation disclosed in US20050234691 describes a method of crop yield prediction which involved utilisation of data of various environmental factors collected by remote sensing technologies. It focused on methods to get accurate predictions by analysing data such as Normalized Difference Vegetation Index (NVDI), soil moisture, soil temperature, and rainfall data. It allowed for giving insight on not just crops like soy and corn but also for other agricultural products such as fruits, vegetables, nuts, etc.
WO2017194276A1 describes a method for detecting plant diseases by using image-based diagnostics and statistical inference methods. It integrates sensors which capture images, processes the image to normalize the colour and then extracts the required parts from the image. Which is then clustered and filtered using Bayesian filters. The disease classification is done by a base and a meta classifier which produces a confidence score for each disease and chooses the highest. It aims to be accurate and efficient.

US11263707B2 describes a way to use machine learning to improve agriculture and agricultural practices. It uses geographical, weather, agronomic and environmental data to predict the crop yield and ideal farming operations. Farmers can request predictions in this method based on the plot of land, location data, weather condition data and soil composition of that area. This aims to help farmers make more informed decisions to improve crop yield prediction.
CN116630803A describes an invention that focuses on tomato plant diseases and insect and pest detection using deep learning. It takes images of plant diseases and insects pests and then a model is trained using YOLOX target detection network. It labels the plant diseases and pest category of the plant disease. Then enhances data and uses a Fisher discrimination CNN network to get the predictions. This utilises two networks to identify diseases and pests category respectively.
AU2021101682A4 presents a method for plant leaf disease diagnosis using machine learning and deep convolutional neural networks. It loads AlexNet and GoogleNet Pretrained Convolutional Neural Networks and configure the last 3 layers such that it tests the result after training by checking its accuracy and improves the result by training the network using that data to increase accuracy.
CN110443420B presents a way to predict crop yield based on machine learning. This approach combines three different machine learning classification algorithms to respectively predict crop yield and its judges the prediction results so that the poor robustness of single learning algorithm is solved and the accuracy and reliability of the prediction results are improved.
US11900560B2 describes a machine-learning-based system that uses high elevation temporal images over a certain time through a crop cycle of a type of crop. It also generates ground level operational data over that time. The model generates the crop yield predictions based on actual data and hypothetical scenarios where parameters are changed. These predictions help in making changes to operational changes in the farm.
CN110751094B describes a model based on the Google Earth Engine remote sensing images and deep learning method in which spatial matching is done on the data and the meteorological department data in the same research period and the geo spatial data using Google Earth Engine. The model predicts annual crop yield by analysing remote sensing and soil data using Keras.
Summary of the Invention
The proposed innovation introduces a Random Forest and YOLOv8 based Crop Yield Prediction and Plant Leaf Disease Detection. This innovation uses leaf images to detect leaf disease and uses crop data like moisture, temperature, soil type etc. to predict the crop yield.

The model is intended to overcome the constraints of existing models as they might not perform well in low light and are not as accurate as YOLOv8 as it can perform real-time leaf disease detection and is robust for leaf disease detection. Random forest combats overfitting and handles non linear relationships between environmental factors well.
In this invention, Random Forest and YOLOv8 work in tandem together to provide a comprehensive solution to agriculture. Random Forest can predict crop yields by analysing environmental and historical data, offering insights into what is expected in terms of productivity. Meanwhile, YOLOv8 detects leaf diseases in real time by detecting and localizing infected areas in plant images. The system combines yield forecasting with disease diagnostics, allowing farmers to make data-driven decisions regarding resource optimization and crop health management for better agricultural outcomes.
Brief Description of Drawings
The invention will be described in detail with reference to the exemplary embodiments shown in the figures wherein:
Figure-1: The Image of Random Forest related to Crop Yield prediction.
Figure-2: Flowchart of the workflow of the system
Figure-3: YOLO Architecture Flowchart
Detailed Description of the Invention
The idea for the invention is utilization of modern deep learning approaches, such as Random Forest and YOLO to increase farm efficiency by detecting plant diseases on leaves and predicting crop production outcomes. Besides disease detection information that is incorporated in decision making process, the objective of developing the crop yield prediction model is also to incorporate disease information that may have affected the yield, or in other words, the system wants to give information that can help minimize or adapt to such losses.
The invention employs a two-pronged approach: diagnostic of diseases and estimate of crop yields. The disease detection component employs the YOLO (You Only Look Once) algorithm, which is an advanced object detection model, to detect and categorise plant leaf diseases from images. The real-time detection of diseases means that the system is able to promptly and correctly identify diseases, giving farmers the requisite information to take necessary action.
The crop yield prediction component employs the Random Forest algorithm, a robust ensemble learning technique, to forecast crop yields depending on the climate, soil type, and previous crop yields. Thus, using disease detection data as an extra input, the model is capable of modifying yield estimates taking into consideration the diseases that may affect the final yield, and therefore produce more accurate predictions.
The YOLO model is trained to detect plants and their diseases using a larger dataset of plant leaf images labelled for diseases. The model’s architecture also enables it to diagnose many diseases in a single image and give a comprehensive report of the plant’s health status. The images are normalized and augmented through the system to improve the quality and standard of the data fed into the system. Real time detection of the crop by YOLO also means that farmers get an instant result on the health of their crops.
Random Forest model uses a wide range of input data such as, historical crop yield data, weather data, soil data and data related to crop disease detection. This is due to the use of multiple decision trees making up the model which makes the model more accurate and less prone to overfitting. Through incorporation of disease detection data, the model can fine tune yield forecasts based on the diseases that affect productivity of crops. This integration enables the system to give better and accurate yield forecasts.
The system integrates the results of the YOLO and Random Forest algorithms to develop a crop yield prediction. The data of disease detection is then incorporated to the yield predictions to come up with a better and realistic yield that takes into account disease losses. Also, the system provides suggestions and advice for disease control and crop treatment in the form of decision support. It also includes suggestions on what pesticides should be applied to the plant depending upon the disease.
The invention seeks to have very high accuracy in identifying the plant leaf diseases and categorizing them so as to enable proper treatment to be provided on time. The inclusion of disease detection data into crop yield prediction model is likely to improve the yield prediction and hence assist farmers in decision making process. The system will seek to offer decision support and advice in the hope of increasing the efficiency of farming and decreasing the number of losses in crops. , Claims:The scope of the invention is defined by the following claims:

Claims:
1. The unified platform for precision farming integrates Random Forest for crop yield prediction and YOLOv8 for leaf disease detection, providing farmers with an efficient and data-driven agricultural tool.,
a) YOLOv8's robust detection capabilities enable accurate identification of diseased leaf regions, even in challenging conditions such as low light, enhancing the reliability of disease diagnosis.
b) Random Forest utilizes environmental data—including water availability, soil quality, rainfall, and temperature—to deliver precise crop yield predictions, aiding in proactive farming strategies.
2. As per claim 2, Random Forest’s ensemble approach reduces the risk of overfitting, ensuring reliable and stable predictions across diverse datasets, even under varying agricultural conditions.
3. As per claim 1, the integrated platform offers actionable insights based on real-time and historical data, helping farmers adapt quickly to changing conditions and plan for future challenges.
4. As per claim 1, the platform is designed for ease of accessibility, enabling farmers to leverage cutting-edge AI tools for sustainable and efficient agriculture.