Optimized Neural Network for Spatial Land cover Classification with the aid of Co occurrence Matrix.

Kushardono, Dony; Fukue, Kiyonari; Shimoda, Haruhisa; Sakata, Toshibumi

doi:10.4287/jsprs.34.4_22

Cited by 5 publications

(5 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This can be seen from the shorter processing time and lower error, and by using test site data, a high classification accuracy is obtained, namely Overall Accuracy of 89% and Kappa Coefficient of 0.88 on the architectural model 1 hidden layer. This is in line with the results of previous studies (Kushardono, Fukue, Shimoda, & Sakata, 1995;Li, Zhang, & Huang, 2022;Silva, Xavier, da Silva, & Santos, 2020) that to get fast convergence in training using backpropagation it takes the number of neurons and the number of layers in the hidden layer that is balanced with the number of neurons in the output layer, in this study the architecture of 1 hidden layer was obtained, with 16 neurons is the best for 10 neurons in the output layer to accommodate 10 class categories with 4 neurons in the input layer.…”

Section: Optimization Of the Neural Network Layersupporting

confidence: 88%

“…The backpropagation-based training process is carried out until the RMS error is lower than 0.01 or a maximum of 10,000 iterations. Based on previous experience, including from research by (Kushardono, Fukue, Shimoda, & Sakata, 1995) that neural network learning with back propagation is successful if there are no errors between the output layer output and the learning teacher RMS error is less than 0.1 and in research it is used 0.01 to further ensure that learning is perfectly convergent. While 10,000 iterations are also based on the research experience that if there are more learning iterations than that it is a failure or the RMS error is still high and cannot converge.Where in this process the weight and bias factors for each neuron in each iteration are corrected based on the RMS error which is calculated from the difference between the output of the neurons in the output layer and the wanted output according to the training data class.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Optimized Artificial Neural Network for the Classification of Urban Environment Comfort using Landsat-8 Remote Sensing Data in Greater Jakarta Area, Indonesia

Sari

Kushardono

Mukhoriyah

et al. 2023

JAETS

View full text Add to dashboard Cite

The development of computer vision technology as a type of artificial intelligence is increasing rapidly in various fields. This method uses deep learning methods based on artificial neural networks, a well-performed algorithm in multi-parameter analysis. One of the development of computer vision models and algorithms is for a thematic digital image classification, such as environmental analysis. Remote sensing based digital image classification is one of the reliable tools for environmental quality analysis. This study aims to perform neural network optimization for the analysis of the urban environment comfort based on satellite data. The input data used are 4 types of geobiophysical indexes as urban environmental comfort parameters derived from cloud-free annual mosaics Landsat-8 remote sensing satellite data. The results obtained in this study indicate that the 1 hidden layer neural network architecture with 16 neurons for the classification of urban environmental comfort and 10 other land cover classes is quite good. The result of the classification using this optimized artificial neural network shows that the distribution of classes is very uncomfortable which dominates the Greater Jakarta area and its surroundings. For other classes in the study area, some are uncomfortable and rather comfortable. By using this method, we obtained a fast classification training time of 18 seconds for 145 iterations to achieve an RMS Error of 0.01, and has a fairly high classification accuracy overall 89% with a Kappa coefficient of 0.88, while the 2 hidden layer neural network architecture does not succeed in achieving convergence

show abstract

Section: Optimization Of the Neural Network Layersupporting

confidence: 88%

Section: Methodsmentioning

confidence: 99%

Optimized Artificial Neural Network for the Classification of Urban Environment Comfort using Landsat-8 Remote Sensing Data in Greater Jakarta Area, Indonesia

Sari

Kushardono

Mukhoriyah

et al. 2023

JAETS

View full text Add to dashboard Cite

show abstract

“…Our ANN classification parameters was chosen based on previous studies conducted by (G. M. Foody & Arora, 1997). Based on Kushardono et al (1995), the optimum results for ANN can be achieved when the numbers of hidden layers kept at minimumum, with hidden neuron layer twice the amount of classes being investigated. The training rate and training momentum should be on the opposite spectrum.…”

Section: Baresoilmentioning

confidence: 99%

“…(G. M. Foody & Arora, 1997) further agreed with previous founding and stating that increasing the size and complexity of the network will incur extra computational and ground data cost, with only a slight increase on classification accuracy. The more hidden layers being used, the more interaction between neurons and in between layers, where the process of finding the offset and gain value of each neuron is more complicated (Kushardono et al, 1995). We, therefore, decided to use only a single hidden layer to optimize computational time and classification accuracy.…”

Section: Baresoilmentioning

confidence: 99%

Assessing the Potential of LAPAN-A3 Data for Landuse/landcover Mapping

Zylshal

Wirawan

Kushardono

2018

IJG

Self Cite

View full text Add to dashboard Cite

LAPAN-A3 / LAPAN-IPB is the third generation of micro-satellite developed by Indonesian National Institute of Aeronautics and Space (LAPAN). The satellite carries a multispectral push-broom sensor that can record the earth's surface at the visible and near-infrared spectrum. Being launched in June 2016, there has no been many publications related to the use of LAPAN-A3 multispectral data for landuse/landcover (LULC) mapping. This paper aims to provide information regarding the use of LAPAN-A3 data for the LULC extraction maximum likelihood algorithm as well as neural network and then evaluate the results. The LAPAN-A3 image was geometrically corrected by using Landsat-8 OLI image as reference data. Three test areas with a size of 1200x945 pixels are then selected for pixel-based classification with the two aforementioned algorithms. For comparison, both LAPAN-A3 and Landsat-8 data were classified for 3 test areas. Accuracy assessment was performed on both datasets using manually interpreted SPOT-6 Pansharpened image as reference data. Preliminary results showed that LAPAN-A3 were able to extract 10 different LULC classes, comprises of built-up area, forest, rivers, fishponds, shrubs, wetland forests, rice fields, sea, agricultural land, and bare soil. The overall accuracy of LAPAN-A3 data is generally lower than Landsat-8, which ranges from 49.76% to 71.74%. These results illustrate the potential of LAPAN-A3 data to derive LULC information. The lack of necessary parameters to perform radiometric correction and blurring effect are several issues that need to be solved to improve the accuracy LULC.

show abstract

“…Tingginya akurasi hasil klasifikasi pada penggunaan model 242 kanal, selain jumlah neuron lapis masukan yang berhubungan dengan neuron lapis tengah menjadi semakin banyak, dimana pada penelitian Kushardono et al (1995a) jumlah neuron pada lapisan ini semakin banyak semakin memudahkan pengklasifikasi neural network dalam membedakan kelas, terutama pada klasifikasi dengan jumlah kelas penutup lahannya yang banyak. Selain itu dengan model 163 kanal yang tidak mempergunakan informasi pada panjang 1235nm hingga 1356nm dan 1721nm hingga 1900nm, sebagaimana nampak pada Gambar 3-1 pada daerah panjang gelombang tersebut masih terdapat sebagian informasi kelas lahan utamanya kawah dan magma, yang kemungkinan informasi ini dijadikan pembeda utama dalam klasifikasi non parametrik ini pada neuron yang terkait, sehingga menyebabkan turunnya akurasi hasil kelas lahan ini menjadi 88% dan 44% dibanding dengan model 242 kanal yang akurasinya dapat mencapai 100% dan 46%.…”

Section: Metodeunclassified

Klasifikasi Penutup/Penggunaan Lahan Dengan Data Satelit Penginderaan Jauh Hiperspektral (Hyperion) Menggunakan Metode Neural Network Tiruan

Kushardono¹

2017

jurnal_inderaja

Self Cite

View full text Add to dashboard Cite

Hyperspectral remote sensing data has numerous spectral information for the land-use/land-cover (LULC) classification, but a large number of hyperspectral band data is becoming a problem in the LULC classification. This research proposes the use of the back propagation neural network for LULC classification with hyperspectral remote sensing data. Neural network used in this study is three layers, in which to test input layer has a number of neurons as many as 242 to process all band data, 163 neurons, and 50 neurons to process the data band has a high average digital number, and data bands at wavelengths of visible to near infrared. The results showed the use of all the data band hyperspectral on classification with the neural network has the highest classification accuracy of up to 98% for 18 LULC class, but it takes a very long time. Selecting a number of bands of precise data for classification with a neural network, in addition to speeding up data processing time, can also provide sufficient accuracy classification results.ABSTRAKData penginderaan jauh hiperspektral memiliki informasi spektral yang sangat banyak untuk klasifikasi penutup/penggunaan lahan (LULC), akan tetapi banyaknya jumlah band data hiperspektral menjadi masalah dalam klasifikasi LULC. Penelitian ini mengusulkan penggunaan back propagation neural network untuk klasifikasi LULC dengan data penginderaan jauh hiperspektral. Neural network yang dipergunakan 3 lapis, dimana untuk uji coba lapis masukan memiliki jumlah neuron sebanyak 242 untuk mengolah seluruh band, 163 neuron, dan 50 neuron untuk mengolah data band yang memiliki nilai digital rataan yang tinggi, dan data band pada panjang gelombang cahaya tampak hingga infra merah dekat. Hasil penelitian menunjukkan penggunaan seluruh band data hiperspektral pada klasifikasi dengan neural network memiliki akurasi hasil klasifikasi tertinggi hingga 98% untuk 18 kelas LULC, akan tetapi waktu yang diperlukan sangat lama. Pemilihan sejumlah band data yang tepat untuk klasifikasi dengan neural network, selain mempercepat waktu pengolahan data, juga bisa memberikan akurasi hasil klasifikasi yang mencukupi.

show abstract

Optimized Neural Network for Spatial Land cover Classification with the aid of Co occurrence Matrix.

Cited by 5 publications

References 6 publications

Optimized Artificial Neural Network for the Classification of Urban Environment Comfort using Landsat-8 Remote Sensing Data in Greater Jakarta Area, Indonesia

Optimized Artificial Neural Network for the Classification of Urban Environment Comfort using Landsat-8 Remote Sensing Data in Greater Jakarta Area, Indonesia

Assessing the Potential of LAPAN-A3 Data for Landuse/landcover Mapping

Klasifikasi Penutup/Penggunaan Lahan Dengan Data Satelit Penginderaan Jauh Hiperspektral (Hyperion) Menggunakan Metode Neural Network Tiruan

Contact Info

Product

Resources

About