Rice disease identification with spectral analysis
Answers
Explanation:
Paddy Diseases Identification with Texture Analysis using Fractal Descriptors Based on Fourier Spectrum Auzi Asfarian, Yeni Herdiyeni Computer Science Department Bogor Agricultural University Bogor Indonesia
Abstract—The efforts to increasing the quantity and quality of rice production are obstructed by the paddy disease. This research attempted to identify the four major paddy diseases in Indonesia (leaf blast, brown spot, bacterial leaf blight, and tungro) using fractal descriptors to analyze the texture of the lesions. The lesion images were extracted manually. The descriptors of ‘S’ component of each lesion images then used in classification process using probabilistic neural networks. This techniques achieved at least 83.00% accuracy when identifying the diseases. This method has a potential to be used as one of the feature if it combined with other features, especially when two diseases with relatively same color involved. Keywords—paddy disease; fractal descriptors; texture analysis.
I. INTRODUCTION The efforts to increase the quantity and quality of rice production to satisfy the increasing needs of rice in Indonesia experienced several obstacles, one of which is the attack of the diseases on paddy fields. Indonesian Directorate General of Food Crops [4] stated that during the period of October 2011 to March 2012, 80,096 hectares of paddy fields exposed to attach by three major paddy diseases in Indonesia: tungro, leaf blast, and leaf blight. To control these diseases and to minimize the impacts of the attacks, the diseases must be identified quickly. Unfortunately, experts who are able to identify the diseases are often unavailable in some region [13]. Computer vision is a potential solution to tackles this problem. One way to identify the diseases in plants is by observing the physical changes (diseases spots or lesions) caused by chemical changes in the sick plants [10]. The images of these spots can be processed and used to recognize the diseases quickly, easily, and inexpensively [13]. This method also nondestructive [2] and the results are consistent. This method involves the extraction the features of the said disease lesion. The common paddy lesion features are the texture, the color, the position, or the size of spots or lesions [7]. Some research combined more than one of these features. For example, [1] used the texture, color, and shape to recognize blast, sheath blight, and brown spot, the three major rice diseases in Sri Lanka, with 70% accuracy. [15] used a color features (e.g. boundary color and spot color) to recognize blast, brown spot, and narrow brown spot diseases and achieved 87.5% accuracy. [6] proposed a new technique to analyze the texture using fractal descriptors based on image Fourier spectrum. When tested to four different datasets (Brodatz, USPTex, OuTex, and plant leaves), this method is more accurate and faster than any other fractal descriptor estimation techniques. This research attempted to identify the four major paddy diseases in Indonesia using fractal descriptors proposed by [6] and assess the performance of said method. The four diseases are leaf blast (Pyricularia orizae), brown spot (Helminthosporium oryzae), bacterial leaf blight (Xanthomonas oryzae), dan tungro (tungro bacilliform virus). Probabilistic Neural Networks (PNN) was chosen as the classifier because its good results in classifying plant diseases [14] and its fast process [11] which is necessary in mobile environment that will be used when the system is ready to implemented.
II. COMMON PADDY DISEASES IN INDONESIA The Directorate General of Food Crops, Ministry of Agriculture of the Republic of Indonesia, routinely monitors some dangerous diseases on paddy crop. Table 1 presents the data on the size of the six major pests attacks on the rice fields. Three diseases that are on the table along with a brown spot disease were used in this study. The sample image of the infected leaves and lesions are presented in Fig 1.