Paper | Year | Objectives | Algorithms | Data used | Results |
[21] | 2012 | Automatic detection and grading of mango fruit according to size, color and skin of the mango. | Fuzzy image clustering algorithm | Dataset of self-collected mango images. | 80% |
[20] | 2013 | Grading of mango, grape and pomegranate affected by anthracnose. | Thresholding, region growing, K-means clustering and watershed + ANN classifier | Dataset of 600 fruits’ image samples collected by a color Digital Camera. | Normal fruits: 84.65% affected fruit: 76.6% |
[18] | 2016 | Automated grading of mango (Mangifera Indica L.) according to maturity and quality. | SVR + MADM | Mango video images captured by a CCD camera. | 87% |
[12] | 2017 | Mango diseases recognition. | K-Means Clustering + MRKT + ANN | 500 images of mango leaves and fruits from a Nikon 16 MP digital camera. | 98% |
[13] | 2017 | Mango diseases recognition. | K-Means Clustering + MRKT + ANN + Wavelet-PCA | 500 images of mango leaves and fruits from a Nikon 16 MP digital camera. | Flower: 98.50%; Leaf: 98.70%; Fruit: 98.75% |
[19] | 2017 | Defect identification and maturity detection of mango fruits. | Proposed image processing algorithms | Dataset of 28 mango images, including 14 defected mangos and 14 did not defected others. | Not specified |
[8] | 2018 | Mango Leaf Ailment (disease) Detection (MLAD). | ANN (Neural Network Ensemble) + SVM | Own dataset. | Mean accuracy of 80% |
[14] | 2018 | Mango leaf unhealthy region detection and classification. | K-means + GLCM and Multiclass SVM | The dataset contains images divided into 5 classes as follows: 61 images for anthracnose, 50 for red rust, 55 for black rot, 75 for scab and 45 for healthy mango leaf. | 96% |
[11] | 2020 | Mango leaf disease detection and identification. | Neural Network Ensemble (SVM) + K-means clustering | Self-collected images. | Average accuracy of 80% |
[16] | 2021 | Mango leaf disease detection and identification. | CCA + cubic SVM | 29 self-collected RGB images using different types of gadgets and from different areas. 135 images after pre-processing. | 95.50% |