Crop leaf images are a great source of information about plant disease and structural behavior, thus they should be retrieved and studied thoroughly. In the identification and study of leaf diseases, image processing is important. The approach used here is crop leaf disease identification is depicted in the diagram below.
(a) Image Acquisition: Diseased leaves are photographed. The images in this collection are in JPEG format which covers a wide range of plant diseases. After that, the images are read.
(b) Image Pre-processing: Various image pre-processing techniques are used to remove noise or exclude other objects from the image. Because the pixel size of the original image is large and the overall operation takes more time, image scaling is used to convert the original image to a thumbnail. After converting the image to a thumbnail, the pixel size will drop and the whole process will take less time.
(c) Image Segmentation: In a targeted application, image segmentation is one of the most commonly used strategies to well separate the pixels of an image. It divides a picture into several discrete states resulting in high similarity between pixels in every region and high dissimilarity between regions.
(d) Feature Extraction: Disease detection highly depends on feature extraction techniques. It is important in determining the identity of an object. Feature extraction is used in a wide variety of image processing applications. Color, texture edge, as well as shape are some of the features used in disease detection.
(e) Detection and Classification of Plant Diseases: Classifiers: Based on the acquired data, classifiers are used to identify and classify several diseases affecting plant leaves. K-Nearest Neighbors (K- NN), Support Vector Machines (SVM), Convolutional Neural Networks (CNN), and Artificial Neural Networks (ANN) are some of the classifiers that have been used in the past to detect diseases in plants.
Comparative Study of Different Classification Techniques for
A. Convolutional Neural Networks: Convolutional neural networks are distinguished from other neural networks by their superior performance with image, speech, or audio signal inputs. The purpose of CNNs is to reduce images to an easy-to-process form without compromising the features that are required to obtain a good prediction. There are different available architectures for CNNs such as Alexnet, Googlenet, VGGNET, etc.
B. Support Vector Machines (SVM): Support Vector Machines, or SVMs, are a popular supervised learning technique that can be used to solve both classification and regression issues. However, it is mostly used in machine learning for classification difficulties. The purpose of the SVM algorithm is to find the optimal line or decision boundary to classify the n-dimensional space into classes so that additional data points can be easily placed into the correct category in the future. A hyperplane is the optimal choice boundary. SVM selects the extreme points/vectors of the hyperplane. The support vectors are the extreme examples, and the algorithm is called a support vector machine. Consider the diagram below, which shows how a decision boundary or hyperplane is used to classify two different groups.
C. K-Nearest Neighbor: The k-nearest neighbor algorithm based on the supervised learning technique and is one of the most basic machine learning algorithms. The NN algorithm assumes that the new case/data but also the existing cases are comparable and also puts the new case in the category that is most similar to the existing category. The KNN algorithm maintains each available data as well as classifies the new data points into groups that are really similar to the previous data. This means that when fresh data comes in, the k-nn algorithm can quickly classify it into an appropriate category. The KNN algorithm is used for both regression as well as classification, but it is more commonly used for classification tasks. The k-NN algorithm is actually a non-parametric algorithm, that means it does not make any assumptions about the data. It is also known as a lazy learner technique because it does not learn from the set immediately; Alternatively, it saves the dataset as well as performs an action on it when it comes time to classify it. During the training phase, the KNN algorithm simply stores the dataset, and then when it receives the dataset, it classifies it into a category that is similar to the new data.
D. Artificial Neural Networks: Biological neural networks establish the functioning of the human brain, and the phrase “artificial neural network” is derived from them as well. Artificial neural networks, like the human brain, have neurons that are coupled to one another in different layers of a network. Nodes are the name for these neurons.
Conclusion
AI can benefit agronomists to increase the amount of production in addition to decrease the cost of production and labor. It is needless to say that the spread of AI popular application areas today is determined to take a perfect change in the method we prepare studies today in addition to expanding in agronomy. AI changes near robotics additionally by means of extra accuracy to achieve real-time management, which normally serves in old-style agriculture towards precision agriculture by means of low cost. AI resolution is available in addition to essential agricultural public. The present review study presents the diverse applications of AI in the agricultural sector. Severe diseases in crops are the main for the annual sufferers of the cultivated crop. Therefore, recognizing the diseases at an early stage in the crop is very important with the aim of anticipating such extreme sufferers. This paper reviewed the monocotyledonous in addition to dicotyledonous plant relationships by means of connected method about the image processing steps. Different techniques which are generally used solely for the identification in addition to the recognition of diseases are reviewed. The analysis of the utmost present work is additionally depicted in the tables. Different techniques which are utilized in the present work, the maximum accuracy is accomplished by means of deep learning ideas which is more than CNN approach. Some challenges in the agricultural disease detection field that have not been resolved are also discussed.
Read Also:
- Study On Artificial Intelligence Applications Uses In Agriculture
- Advantages And Disadvantages Of Robots In Agriculture
- Role Of Robots In Agriculture
- Artificial Intelligence In Agriculture
- Importance Of Artificial Intelligence And Machine Learning In Agriculture
- Artificial Intelligence (AI) In Agriculture: Current Status And Future Need
- A Glimpse About Artificial Intelligence (AI) In Agriculture
Leave a Reply