difference between early blight and late blight

The overall detection result was 96.2% based on spectral information divergence (SID) classification method. 14 discriminated citrus canker disease from healthy, insect damaged, greasy spot, melanose, scab and wind scar infected samples using the hyperspectral imaging technique. However, no texture features were studied for the above researches. The detection result for infected sugar beet leaves was 97.0%. Vegetation indices were related to different physiological parameters and SPAD could reflect the chlorophyll content. The eight spectral vegetation indices and soil and plant analyzer development (SPAD) were used as features for detection. 13 detected and classified diseased sugar beet leaves based on support vector machine (SVM) classifier model and spectral vegetation indices. Spectral angle mapper (SAM) and head blight index (HBI) methods obtained correct detection rate of 84.0% and 91.0%, with an error of ☑0.0%. PCA was used to identify four distinct wavelength ranges (500–533 nm, 560–675 nm, 682–733 nm and 927–931 nm), which allow the classification of healthy and diseased ear tissues. 12 used hyperspectral imaging to detect Fusarium culmorum on wheat ears. A partial least squares (PLS) regression model was established to predict the infection severities. In that study, principal component analysis (PCA) was used to remove background, bad pixels and shading of the images. 11 investigated fungal (Fusarium verticillioides) development in maize kernels using hyperspectral imaging technique. This technique has been studied previously for identification of diseases on crops. Each pixel for one hyperspectral image has wavelengths covering the full spectral range. Hyperspectral imaging, which covers both spectral and imaging information, has been widely used in many fields, such as food 6, agriculture 7, medical science 8, geography 9 and archaeology 10.The hyperspectral image (a hyperspectral cube) is composed of a series of images covering the whole wavelengths. Also, highly trained and qualified technicians are required for such detection techniques. However, these methods have shown to be time-consuming, inefficient and destructive. Diagnose testing methods include polymerase chain reaction (PCR), enzyme linked immune sorbent assay (ELISA), fluorescence in situ hybridization and biomarker-based detection technology 5. When these fungi infect tomato leaves, symptoms can rapidly spread and cover the entire leaf blade in a favorable environmental condition.Ĭonventionally, detecting diseases on crop leaves are based on visual assessments and diagnostic methods. Early blight ( Alternariasolani) and late blight ( Phytophthorainfestans) are the two common fungal diseases of tomato plants 4. However, the quality and yield of tomato can be easily affected by various diseases during the growing season.

They have anticarcinogenic, cardioprotective and other health benefits 3. Tomatoes are a rich source of lycopene, beta-carotene, folate, potassium, vitamin C, flavonoids and vitamin E 2. The results demonstrated that hyperspectral imaging has the potential as a non-invasive method to identify early blight and late blight diseases on tomato leaves.Īccording to a previous research, tomato fruit ranks the first place among 40 fruits and vegetables in “relative contribution to human nutrition” 1 and is widely consumed due to its high nutrition value. Among the eight texture features, dissimilarity, second moment and entropy carried most of the effective information with the classification accuracy of 71.8%, 70.9% and 69.9% in the ELM models. Among the models which were established based on spectral information, all performed excellently with the overall classification accuracy ranging from 97.1% to 100% in testing sets. Then, eight texture features (mean, variance, homogeneity, contrast, dissimilarity, entropy, second moment and correlation) based on gray level co-occurrence matrix (GLCM) at the five effective wavelengths were extracted to establish detection models. Based on the five selected wavelengths (442, 508, 573, 696 and 715 nm), an ELM model was re-established. Successive projections algorithm (SPA) was used to identify the most important wavelengths. An extreme learning machine (ELM) classifier model was established based on full wavelengths. One hundred and twenty healthy, one hundred and twenty early blight and seventy late blight diseased leaves were selected to obtain hyperspectral images covering spectral wavelengths from 380 to 1023 nm. This study investigated the potential of using hyperspectral imaging for detecting different diseases on tomato leaves.

Difference between early blight and late blight