Solid Waste Bin Classification using Gabor Wavelet Transform
Arivienanand Rajamanikam1, Mahmud Iwan Solihin2
1Arivienanand Rajamanikam, Faculty of Engineering, Technology & Built Environment, UCSI University, Jalan Menara Gading, Lumpur, Malaysia.
2Mahmud Iwan Solihin, Faculty of Engineering, Technology & Built Environment, UCSI University, Jalan Menara Gading, Lumpur, Malaysia.
Manuscript received on 01 February 2019 | Revised Manuscript received on 07 February 2019 | Manuscript Published on 13 February 2019 | PP: 114-117 | Volume-8 Issue- 4S February 2019 | Retrieval Number: DS2846028419/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Solid waste dumping has become an issue and a threat towards health and it is continuing to deteriorate with time. It is known that solid waste (SW) accumulation will rise drastically over time and epidemiological effects will soon rise from unplanned or unscheduled solid waste dumping. With Solid Waste Management (SWM) at its optimum performance, this problem can be mitigated. There are several reasons for a system to fail but one is focussed to engineer a solution towards waste treatment. Solid waste segregation takes longer to process compared to treating it on a weekly schedule. By utilising machine vision and machine learning technologies, solid waste bin classification can be done norm as a pathway towards efficient waste segregation. In this paper Gabor wavelet transformation (GWT) is used for classifying solid waste images by convoluting an image with Gabor wavelet kernels with different scales and orientation. The features are extracted from the image training database to model a supervised Artificial Neural Network (ANN) with the actual bin level grades. The computational speed or efficiency of the GWT is increased by using Genetic algorithm (GA) where a total of 48 out 80 features are used sufficiently, whereby less wavelets are used in the process thus increasing the performance to a maximum of 47.52%. The mean squared error before and after optimisation gave a difference of 91.9% in improvement with GA. The proposed method proved that with GWT and GA, SW is gradable with random waste images and it has proven to be optimum from analysis.
Keywords: Gabor Wavelet Transformation (GWT), Solid Waste Management (SWM, solid waste (SW), Artificial Neural Network (ANN).
Scope of the Article: Classification