International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 08 | Aug 2024
p-ISSN: 2395-0072
www.irjet.net
CITRUS FRUIT DISEASES RECOGNITION BASED ON SINGLE STREAM MODIFIED MOBILENET V2 MODEL 1
2
A.RAJAGOLDINA , Ms. NISHA G KRISHNAN., M.E.,
1PG Student, Dept. of Electronics and Communication Engineering, Rohini College of Engineering and
Technology, Kanniyakumari Tamil Nadu, India 2Assistant professor, Dept. of Electronics and Communication Engineering, Rohini College of Engineering and
Technology, Kanniyakumari, Tamil Nadu, India. -------------------------------------------------------------------------***------------------------------------------------------------------------------utilizes progressed man-made intelligence innovations to Abstract — Agriculture is a huge examination field in
gather better yields, control bug, notice soil and developing circumstances, sort out information for ranchers, help with responsibility, and making improvement a few farming based processes for the total food inventory network. As of late, accuracy agribusiness utilizes smart methods to make due, comprehend, and coordinate every one of the "enormous information" being created from the homestead. Fortunately, computer algorithms are used in machine learning (ML) techniques to parse data, learn from it, and make decisions in order to identify disorders, diseases, and pests. Beginning around 2012, PC vision (CV) and profound learning (DL) models have begun to get comfortable in different fields of picture handling, object limitation and acknowledgment in the pictures. In this view, this exploration work is focused on the ID and characterization of citrus illnesses utilizing ML and DL models.
the field of plant picture handling. The production of agricultural goods is essential to the expansion of the national economy and the global economy. Horticulture likewise essentially affects a country's capacity to beat destitution. The recognition of fruit diseases is quickly becoming a popular topic in computer vision. The presence of plant illnesses decreases organic product creation as well as purposes a huge misfortune to the public economy. Citrus organic products help to fortify the insusceptible framework, permitting it to fend off sicknesses like Coronavirus. Manual investigation of natural product illnesses with the unaided eye takes time and is troublesome; thusly, a PC based technique is constantly expected for exact acknowledgment of plant infections. In existing framework two-stream profound learning procedure is utilized. It takes a long time to compute. To conquer the downsides, in this task, single stream convolutional brain network engineering is proposed for perceiving citrus organic product sicknesses. Four contrast enhancement operations are used for data augmentation in the first step: shadow expulsion, changing pixel power, further developing splendor, and working on neighborhood contrast. In the second step, the MobileNet-V2 CNN model is chosen and adjusted. The fine-tuned model is trained on the augmented citrus dataset through transfer learning. The recently prepared model is utilized for profound component extraction; However, analysis reveals that the extracted deep features contain very little information that is redundant.
As of now, the majority of individuals don't pick the farming area and there is an extreme easing which brings about complex issues. Basically, in order to harvest crops and make a profit, farmers and farming lands need more workers. Unfortunately, people these days live in cities. way of life and cultivating has been a dull space for the majority of the people. A significant arrangement is computer based intelligence horticulture bots, which expand the human work labor force which can be applied at different cases. Likewise, it is utilized in collecting crops at a most extreme amount and fast speed when contrasted and human works. In addition, it helps farms save money on regular labor by accurately locating weeds and bugs. In the agricultural sector, AI techniques can assist farmers in a variety of ways. With the help of Al, farmers can use the collected farm data to examine real-time parameters like temperature, water level, and soil conditions to make better decisions..
Key Words: CNN, MobileNet-V2, Machine learning (ML), computer vision (CV). 1. INTRODUCTION Artificial intelligence (A turns out to be more normal in electronic gadgets, which is material in a few spaces going from homes, horticulture, training, diversion, business, web based business, etc. Universally, horticulture is a $5 trillion industry and it as of late
© 2024, IRJET
|
Impact Factor value: 8.226
M. Nikhithaet al. [1] proposed In India, crop yield is declined because of the post-acknowledgment of illnesses in natural products/vegetables by the ranchers. Ranchers face
|
ISO 9001:2008 Certified Journal
|
Page 183