Rice plant-hopper infestation detection and classification algorithms based on fractal dimension values and fuzzy C-means

Abstract

a b s t r a c tRice plant-hopper (RPH) (Nilaparvata lugens, Sogatella furcifera, and Laodelphax striatellus) infestation is considered one of the most serious disasters in rice production in Asia. In order to use visible images to detect stress in rice production caused by RPH infestation, an algorithm based on fractal eigenvalues and fuzzy C-means (FCM) has been developed. First, an experiment was designed and many visible images of rice stems were collected. Based on the pretreatment of these visible images, such as smoothing, denoising, image color space and frequency domain transformation, the relative image fractal eigenvalues were extracted by using a box-counting dimension method, and then the linear regression model and FCM clustering algorithm were developed. Results showed that it is possible to distinguish RPH gathering areas and mouldy gray spots after RPH damage on rice plant stems from healthy rice plant stems using fractal-dimension values. Most of the fractal eigenvalues based on different-sized subregional image blocks have higher correlation coefficient with the RPH quantity, and the D skewness (skewness), D mean (mean) and D moment (moment), (where D j is the fractal-dimension value of the j-th subregional image block), based on a 16 × 16 pixel subregional image block, have higher correlation coefficients: 0.803, 0.794, and 0.799 respectively. Using RPH infestation detection and classification based on fractal eigenvalues and the FCM algorithm, the accuracy to differentiate cases in which RPH infestation had occurred or not reached 87%. The accuracy to differentiate four groups was 63.5%. The result is expected to meet the needs of actual rice production: along with the use of a micro-sensor mobile monitoring platform, the visible-imagebased method based on fractal eigenvalues and the FCM algorithm can be used to obtain redundancy information for the large area of remote sensing to the stress induced by RPHs, and then to improve the monitoring accuracy in fusion decisions of integrated pest management.

Automatic Airport Recognition Based on Saliency Detection and Semantic Information

“…In the experiment, several state of the art methods are chosen for comparison. They are BOVW, the probabilistic latent semantic analysis (PLSA) method [41], a fractal fuzzy C means method adapted from [42] and the traditional recognition methods based on image matching via SIFT and speeded up robust features (SURF) [43].…”

Recent Studies of Image and Soft Computing Techniques for Plant Disease Recognition and Classification

“…[Zhou et al 2013] [41] proposed an algorithm to detect and classify rice hopper infection (RHP). The infected rice stem images acquired and then applied smoothing, denoising, color space transformation.…”

Classification of lignocellulosic biomass by weighted‐covariance factor fuzzy C‐means clustering of mid‐infrared and near‐infrared spectra

“…each spectrum belongs to each cluster with a certain membership degree . This algorithm is less sensitive to random initialization than K‐means and is therefore more robust and has been successfully used to study agricultural and microbial processes . The classical FCM algorithm is based on Euclidean distance, which is adapted for spherical structural clusters.…”

“…20 This algorithm is less sensitive to random initialization than K-means and is therefore more robust and has been successfully used to study agricultural and microbial processes. 21,22 The classical FCM algorithm is based on Euclidean distance, which is adapted for spherical structural clusters. However, IR spectra are not typically distributed spherically in the data space and are affected by the application of pretreatment methods.…”