Abstract
Plant recognition and diseases identification have an impact on the sustainable development of many countries in theagricultural sector. The automatic plant recognition and diseases identification will assist the specialists and expertsin agriculture to overcome many of plant diseases and problems. The automation of plant diseases identification andrecognition approaches have received considerable interest in the last years because their effect on the growth of theeconomy of countries, which may depend mainly on agriculture and to reduce the economic losses in the sustainableagriculture industry in general. However, human cognition and sight are not sufficient to identify the region of interestin the images of plants, usually, stems and leaves. Nowadays, image-based methods are considered as a visual assistingof plant recognition and diseases identification with the aid of the recent advances in image processing area. In thispaper, we describe and analyze the automated image-based methods and discuss the state-of-art of plant recognitionand diseases identification that has been applied in the last years. Also, we explore the role of image processingmethods and classifiers in plant diseases identification and recognition. Different types of datasets of plant diseasesidentification and recognition are introduced briefly with their existing problems. As an example, the preprocessingphase of this issue is implemented based on real infected tomato leaves. Also, shape feature, color feature, and texturefeature have been reviewed. Moreover, we described the important classifiers that are used currently used in theclassification process. Also, hybrid classifiers can integrate the results from multiple algorithms with the aim ofimproving classification accuracy. Therefore, some of the well-known hybrid classifiers for plant diseasesidentification and recognition have been presented. Some solutions of using image-based methods such as complexbackgrounds of the region of interest, different plant diseases can produce similar symptoms, and the conditions ofcapturing images have been presented. Finally, some points of the future work are proposed.
Highlights
Background removalIn the background of images, a shadow may disturb the feature extraction phase
Shape Feature Shape feature is considered as an important feature in plant images, and the extraction process of shape features usually depends on the segmentation of the image
Artificial Neural Network (ANN) is used in leaf recognition with Probabilistic Neural Network (PNN) classifier for different classes of plant leaves samples, and the results showed high recognition rate of 93.08 % [57]. 2. k-Nearest Neighbor Classifier (KNN) KNN is data mining classification techniques which is based on the closest training examples in the feature space can be represented by its closest K neighbors
Summary
Correlogram (ACC) [19], image retrieval method based on border/interior pixel classification [20], color indexing using GCH (Global Color Histogram) [21]. 3.3. An expert system of plant identification has been proposed in [23] to identify different species of plant based on their images of leaves. In this system, the ant colony optimization algorithm is used for feature selection. The used images in this phase are captured to the infected leaves of tomato with two diseases, Powdery mildew, and early blight, from different farms. This real data set includes 200 infected leaves images were 100 for each virus type has been extracted. Captured images are with different sizes and these images resized to 512x512 resolution to reduce the size
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
