| Autor: |
Sharma, Sudhir, Chhimwal, Nitin, Bhatt, Kaushal Kishor, Sharma, Abhay Kumar, Mishra, Prashant, Sinha, Swati, Raj, Sundeep, Tripathi, Sandesh |
| Předmět: |
|
| Zdroj: |
Wireless Networks (10220038); Oct2021, Vol. 27 Issue 7, p4929-4947, 19p |
| Abstrakt: |
Automatic weed classification is necessary to enhance the productivity of crops in weed management system. Major issue in agriculture is the control of weeds growing among plantation crops. Various site-specific weed management techniques are developed to classify the weeds, but enhancing the production quality and yield poses a complex task in agriculture. An effective method is developed to classify the crop or weeds using the proposed Feedback Cat Swarm Optimization-based Deep Neuro-Fuzzy Network (FCSO-based DNFN). However, the proposed FCSO is derived by integrating the Feedback Artificial Tree (FAT) algorithm and Cat Swarm Optimization (CSO) algorithm. The process of weed classification is carried out with the MapReduce framework by employing the mapper and reducer functions. The cluster head (CH) selection is made, and the process of routing is carried out through CH to transfer the weed image from the sensor nodes to Base Station (BS) for weed classification. However, the classification process is done based on the fitness measure by considering the feature vector. Through the comparative analysis of the proposed system with the existing state-of-art techniques, the proposed FCSO-based DNFN obtained higher performance in terms ofthroughput, energy, delay, and accuracy with the values of 0.5018, 0.5014 J, 0.5174 s, and 0.9523, respectively. Besides, while increasing the node size, the accuracy of the classification improves, and hence the performance of the developed model also increases. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
Full text from SpringerLink