| CPC G06T 7/0012 (2013.01) [G06F 18/23213 (2023.01); G06T 7/11 (2017.01); G06T 7/194 (2017.01); G06T 7/62 (2017.01); G06V 10/25 (2022.01); G06V 10/763 (2022.01); G06V 20/188 (2022.01); G06V 20/20 (2022.01); H04N 7/183 (2013.01); G06T 2207/10024 (2013.01); G06T 2207/10028 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/30188 (2013.01)] | 19 Claims |
|
1. An image-processing based advisory system (100) for precision agriculture and for quality evaluation and sorting of agricultural products, said system (100) comprising:
a. a user device (102) comprising:
i. at least one red green blue (RGB) imaging unit (106) for capturing at least one digital image of a scene, said imaging unit (106) integrated with a first set of sensors (108) to ensure capturing of a digital image under different light conditions;
ii. a processing unit (110) configured to cooperate with said imaging unit (106) to receive said digital image, and further configured to cooperate with a second set of sensors (134) to receive a sensed data corresponding to a pre-determined set of scene-related and environmental parameters;
iii. a first communication module (112) configured to cooperate with said processing unit (110) to receive and transmit said digital image and said sensed data;
iv. a battery (114) for supplying power to at least said imaging unit (106), said first set of sensors (108), said processing unit (110), and said first communication module (112),
b. a cloud server (104) comprising:
i. a second communication module (122) configured to receive said digital image and said sensed data from said user device (102) via a wireless communication network;
ii. a database (124) configured to store a) chemical signature based machine learning and deep learning library dataset and b)spectral signature based machine learning and deep learning library datasets comprising a prior acquired data for different crops and diseases, said data associated at least with abiotic stress symptoms, nutrient deficiency, toxicity symptoms, crop growth stages, growth-stage wise nutrient requirement, information on weeds, and pre-harvest and post-harvest crop quality;
iii. a correlation module (126) configured to cooperate with said database (124) to receive said datasets to train one or more prediction models, said correlation module (126) further configured to construct a three-dimensional HyperintelliStack data structure from said datasets, said HyperintelliStack data structure providing correlations between at least a set of red green blue (RGB) pixel values and hyperspectral reflectance values corresponding to each of said RGB values, each face of said HyperintelliStack data structure representing one primary RGB reflectance and/or value, each of said faces divided into a plurality of cells, wherein each cell provides a pre-trained hyperspectral signature for a given set of RGB values;
iv. a transforming unit (128) configured to cooperate with said correlation module (126) to transform said received digital image made of multiple RGB pixel values into a hyperspectral image using said HyperintelliStack data structure;
v. a computation module (130) configured to cooperate with said transforming unit (128) to compute a plurality of vegetation indices for each pixel of said hyperspectral image, and further configured to generate a segmented image from said received hyperspectral image based on said computed vegetation indices; and
vi. a prediction engine (132) configured to cooperate with said computation module (130) to receive said segmented image, and further configured to cooperate with said correlation module (126) to generate at least one advisory for precision agriculture and for quality evaluation and sorting of agricultural products using said segmented image and said one or more prediction models.
|