AromaticFlowersBD: An Extensive Image Dataset of Aromatic Flowers in Bangladesh

Type of data: 1024x1024 px images of aromatic flowers.
Data format: JEPG

Dataset Contents: Original and augmented images of different varieties of aromatic flowers in Bangladesh from bulk flowers and single flower perspectives.

Number of classes: Twelve common aromatic flower varieties - (1) Bakul, (2) Beli, (3) Champa, (4) Golap, (5) Hasna Hena, (6) Jui, (7) Kath Golap, (8) Madhobi Lata, (9) Rajani Gandha, (10) Sheuli, (11) Sondha Maloti, and (12 Tagar.

Total number of images: (A) In Original Dataset: 5,775. (B) In Augmented Dataset: 51,975.

Distribution of instances:
(A) Number of images in Original Dataset (Aromatic Flowers) - (1) Bakul = 409 images in total. Single = 207, Bulk = 202. (2) Beli = 572 images in total. Single = 284, Bulk = 288. (3) Champa = 507 images in total. Single = 209, Bulk = 298. (4) Golap =452 total images. Single = 214, Bulk = 238. (5) Hasna Hena = 462 images in total. Single = 200, Bulk = 262. (6) Jui = 444 images in total. Single = 244, Bulk = 200. (7) Kath Golap = 438 images in total. Single = 237, Bulk = 201. (8) Madhobi Lata = 516 images in total. Single = 273, Bulk = 243. (9) Rajani Gandha = 504 images in total. Single = 266, Bulk = 238. (10) Sheuli = 467 images in total. Single = 221, Bulk = 246. (11) Sondha Maloti = 494 images in total. Single = 217, Bulk = 277. (12) Tagar = 510 images in total. Single = 223, Bulk = 287.
(B) Number of images in Augmented Dataset (Augmented Aromatic Flowers) - (1) Bakul = 3,681 images in total. Single = 1,863, Bulk = 1,818. (2) Beli = 5,148 images in total. Single = 2,556, Bulk = 2,592. (3) Champa = 4,563 images in total. Single = 1,881, Bulk = 2,682.
(4) Golap = 4,068 images in total. Single = 1,926, Bulk = 2,142. (5) Hasna Hena = 4,158 images in total. Single = 1,800, Bulk = 2,358.
(6) Jui = 3,996 images in total. Single = 2,196, Bulk = 1,800. (7) Kath Golap = 3,942 images in total. Single = 2,133, Bulk = 1,809.
(8) Madhobi Lata = 4,644 images in total. Single = 2,457, Bulk = 2,187. (9) Rajani Gandha = 4,536 images in total. Single = 2,394, Bulk =2,142. (10) Sheuli = 4,203 images in total. Single = 1,989, Bulk = 2,214. (11) Sondha Maloti = 4,446 images in total. Single = 1,953, Bulk = 2,493. (12) Tagar = 4,590 images in total. Single = 2,007, Bulk = 2,583.

Dataset size: (A) Total size of the Original Dataset is 700 MB and the compressed ZIP file size is 682 MB. (B) Total size of the Augmented Dataset is 7.17 GB and the compressed ZIP file size is 7.00 GB.

Data acquisition process: Images of aromatic flowers are captured using high-definition smartphone cameras from different angles and from the two perspectives of single-flower and bulk-flower.
Data source location: Local gardens, plant nurseries, and flower shops located in different areas of Dhaka, Bangladesh.

Where applicable: Training and evaluating machine learning and deep learning models to distinguish aromatic flower varieties in Bangladesh to support automated identification systems of various aromatic flowers.

数据类型：1024x1024 像素的芳香花卉图像。
数据格式：JPEG
数据集内容：收录了孟加拉国不同品种芳香花卉的原版及增强图像，涵盖批量花卉与单株花卉的视角。
类别数量：十二种常见的芳香花卉品种，包括：（1）白兰花（Bakul）、（2）红茉莉（Beli）、（3）依兰（Champa）、（4）茉莉（Golap）、（5）哈斯娜·亨娜（Hasna Hena）、（6）聚雨（Jui）、（7）卡瑟·戈拉普（Kath Golap）、（8）马达胡比·拉塔（Madhobi Lata）、（9）拉贾尼·甘达（Rajani Gandha）、（10）谢丽（Sheuli）、（11）松达·马洛蒂（Sondha Maloti）和（12）塔加尔（Tagar）。
图像总数：（A）原始数据集中总计5,775张图像。（B）增强数据集中总计51,975张图像。
实例分布：（A）原始数据集中芳香花卉图像数量分布——（1）白兰花总计409张，单株207张，批量202张。（2）红茉莉总计572张，单株284张，批量288张。（3）依兰总计507张，单株209张，批量298张。（4）茉莉总计452张，单株214张，批量238张。（5）哈斯娜·亨娜总计462张，单株200张，批量262张。（6）聚雨总计444张，单株244张，批量200张。（7）卡瑟·戈拉普总计438张，单株237张，批量201张。（8）马达胡比·拉塔总计516张，单株273张，批量243张。（9）拉贾尼·甘达总计504张，单株266张，批量238张。（10）谢丽总计467张，单株221张，批量246张。（11）松达·马洛蒂总计494张，单株217张，批量277张。（12）塔加尔总计510张，单株223张，批量287张。（B）增强数据集中增强芳香花卉图像数量分布——（1）白兰花总计3,681张，单株1,863张，批量1,818张。（2）红茉莉总计5,148张，单株2,556张，批量2,592张。（3）依兰总计4,563张，单株1,881张，批量2,682张。（4）茉莉总计4,068张，单株1,926张，批量2,142张。（5）哈斯娜·亨娜总计4,158张，单株1,800张，批量2,358张。（6）聚雨总计3,996张，单株2,196张，批量1,800张。（7）卡瑟·戈拉普总计3,942张，单株2,133张，批量1,809张。（8）马达胡比·拉塔总计4,644张，单株2,457张，批量2,187张。（9）拉贾尼·甘达总计4,536张，单株2,394张，批量2,142张。（10）谢丽总计4,203张，单株1,989张，批量2,214张。（11）松达·马洛蒂总计4,446张，单株1,953张，批量2,493张。（12）塔加尔总计4,590张，单株2,007张，批量2,583张。
数据集大小：（A）原始数据集总大小为700 MB，压缩ZIP文件大小为682 MB。（B）增强数据集总大小为7.17 GB，压缩ZIP文件大小为7.00 GB。
数据采集过程：利用高清智能手机从不同角度拍摄芳香花卉图像，并从单株花卉和批量花卉两种视角进行采集。
数据来源地：孟加拉国达卡市不同区域的本地花园、植物苗圃和花店。
应用场景：用于训练和评估机器学习及深度学习模型，以区分孟加拉国的芳香花卉品种，支持各类芳香花卉的自动识别系统。