Recognition of Rice Sheath Blight Based on a Backpropagation Neural Network

Rice sheath blight is one of the main diseases in rice production. The traditional detection method, which needs manual recognition, is usually inefficient and slow. In this study, a recognition method for identifying rice sheath blight based on a backpropagation (BP) neural network is posed. Firstly, the sample image is smoothed by median filtering and histogram equalization, and the edge of the lesion is segmented using a Sobel operator, which largely reduces the background information and significantly improves the image quality. Then, the corresponding feature parameters of the image are extracted based on color and texture features. Finally, a BP neural network is built for training and testing with excellent tunability and easy optimization. The results demonstrate that when the number of hidden layer nodes is set to 90, the recognition accuracy of the BP neural network can reach up to 85.8%. Based on the color and texture features of the rice sheath blight image, the recognition algorithm constructed with a BP neural network has high accuracy and can effectively make up for the deficiency of manual recognition.

Classification of Computer Graphic Images and Photographic Images Based on Fusion of Color and Texture Features

With the rapid advancement in digital image rendering techniques, allows the user to create surrealistic computer graphic (CG) images which are hard to distinguish from photographs captured by digital cameras. In this paper, classification of CG images and photographic (PG) images based on fusion of global features is presented. Color and texture of an image represents global features. Texture feature descriptors such as gray level co-occurrence matrix (GLCM) and local binary pattern (LBP) are considered. Different combinations of these global features are investigated on various datasets. Experimental results show that, fusion of color and texture features subset can achieve best classification results over other feature combinations.

Color moment and gray level co-occurrence matrix in classification of soil organic matter for patchouli plantation

<span>Patchouli is one of the essential plants that have the most potential and widely cultivated in Indonesia. Patchouli is greedily absorbing soil nutrients and organic matter. Therefore, the selection of soil with high organic matter will maximize the patchouli’s productivity. This paper aims to facilitate soil’s organic matter identification by classifying soil image based on the combination of color and texture features. The color feature extraction was done using the Color Moments method and the texture feature was done using Gray Level Co-occurrence Matrix (GLCM) method. The selection of features was performed to obtain the best combination of color and texture features. The selected features then was used as input of classification by using Modified K-Nearest Neighbor (MKNN). The samples of soil that used as data were taken from several districts in Blitar, East Java province. The testing result of this research showed the highest accuracy of 93,33% by using 180 training data, and also particular color and texture feature combination.</span>

Detection of Drechslera avenae (Eidam) Sharif [Helminthosporium avenae (Eidam)] in Black Oat Seeds (Avena strigosa Schreb) Using Multispectral Imaging

Conventional methods for detecting seed-borne fungi are laborious and time-consuming, requiring specialized analysts for characterization of pathogenic fungi on seed. Multispectral imaging (MSI) combined with machine vision was used as an alternative method to detect Drechslera avenae (Eidam) Sharif [Helminthosporium avenae (Eidam)] in black oat seeds (Avena strigosa Schreb). The seeds were inoculated with Drechslera avenae (D. avenae) and then incubated for 24, 72 and 120 h. Multispectral images of non-infested and infested seeds were acquired at 19 wavelengths within the spectral range of 365 to 970 nm. A classification model based on linear discriminant analysis (LDA) was created using reflectance, color, and texture features of the seed images. The model developed showed high performance of MSI in detecting D. avenae in black oat seeds, particularly using color and texture features from seeds incubated for 120 h, with an accuracy of 0.86 in independent validation. The high precision of the classifier showed that the method using images captured in the Ultraviolet A region (365 nm) could be easily used to classify black oat seeds according to their health status, and results can be achieved more rapidly and effectively compared to conventional methods.