Okra DiseaseNet Dataset

Crop Name : Okra or Lady’s finger (Scientific Name : Abelmoschus esculentus)
Dataset : Okra leaf disease image dataset
Problem : To address the segmentation and classification problems in the field of AI researchers due to data scarcity
in Indian agricultural field.
Motivation : To diagnose the okra leaf diseases segmentation with severity level estimation and classification at
early stage using computer vision algorithms
Dataset Description : This dataset comprises of 1500 high resolution okra leaf healthy and diseased images and also
mainly focus on collection of various types of leaf diseases according to two locations and nature of soil
(Chengalpattu and Thanjavur) includes 1500 images with 6 different classes namely Class 0 : Healthy, Class 1 :
Alternaria leaf spot, Class 2: Cercospora leaf spot, Class 3: Downy mildew, Class 4: Leaf curly virus and Class 5:
Phyllosticta leaf spot.
Segmentation : It will be also used in segmentation to measure the severity level estimation for exactly localize and
segments the diseased lesion spots in each leaf.
Classification : It used to classify the images as healthy and different types of diseased okra leaves.
Image Acquired by : Canon EOS 3000D DSLR Camera with specifications, 18-55 mm lens (Black), 9 Auto focus
points, Wifi, Full HD, Optimal zoom 35x, CMOS Sensor type and 18MP effective pixels.
Image type : RGB
Image Format : jpg format
Image size : 224x224x3
Dataset Split Ratio : Training - 75%, Testing - 15%and Validation -10%
Time Span for data acquisition : Chengalpattu -> March and April, 2023 and Thanjavur -> May, 2023
Location : SRM Urban Farm, Chengalpattu and Agricultural Lands, Thanjavur
Why dataset are needed :
 Okra dataset is the first crop acquired for segmenting and classifying their diseases in Indian agricultural lands
with different light illumination.
 It mainly focus on aiding the farmers to diagnose leaf diseases at early stage and prevent food insecurity issues
for avoiding crop yield losses.
 It also aids to the researchers also to build AI models for superior efficient results.
Citation:
If you find this dataset helpful and use it in your work, kindly cite this dataset using “Kumarakuru, Sowmiya; M, Thenmozhi (2024), “Okra DiseaseNet Dataset”, Mendeley Data, V1, doi: 10.17632/nh7zk4hv8z.1
”