Spectral Weed Detection and Precise Spraying Laboratory of AgroMachinery and Processing Els Vrindts, Dimitrios Moshou, Jan Reumers Herman Ramon, Josse De Baerdemaeker tholieke Universiteit Research sponsored by IWT and the Belgian Ministry of Small Trade and Agriculture
46
Embed
Spectral Weed Detection and Precise Spraying Laboratory of AgroMachinery and Processing Els Vrindts, Dimitrios Moshou, Jan Reumers Herman Ramon, Josse.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Spectral Weed Detectionand Precise Spraying Laboratory of AgroMachinery and
ProcessingEls Vrindts, Dimitrios Moshou, Jan ReumersHerman Ramon, Josse De Baerdemaeker
Katholieke Universiteit
Research sponsored by IWT and the Belgian Ministry of Small Trade and Agriculture
Katholieke Universiteit Leuven
OverviewSpectral measurements of crops and weeds
in laboratoryin field
Processing of spectral data with neural networks Precise spraying
Katholieke Universiteit Leuven
Optical detection of weeds
Techniquesred/NIR detectors (vegetation index)image processing (color, texture, shape)remote sensing of weed patchesreflection in visible & NIR light
different detection possibilities, different scales
Requirements for on-line weed detection:fast & accurate weed detectionsynchronized with treatment
Spectral analysisstepwise selection of discriminant wavelengths multivariate discriminant analysis, based on reflectance response at selected wavelengths (dataset a)
assuming multivariate normal distributionquadratic discriminant rule
classes with different covariance structure
testing the discriminant function: classification of spectra from dataset b
MLP: Multi-Layer Perceptron, PNN: Probabilistic N Network, SOM: Self-Organizing Map, LVQ: Learning Vector Qantization, LLM: Local Linear Mapping
Comparison between methodsCrop-weed classificationCrop-weed classification
Katholieke Universiteit Leuven
PNN MLP SOM LVQ LLM SOM Sugar beet 91 96 88 91 98 R. repens 55 51 49 55 61 C. arvense 74 72 71 74 80 S. arvensis 70 64 63 74 83 S. media 73 70 69 73 71 T. officinale 66 47 58 66 72 P. annua 66 70 61 66 66 P. persicaria 68 79 60 68 76 U. dioica 48 56 48 48 55 O. europaea 94 98 90 94 99 M. lupulina 87 92 84 87 93
Crop-weed classificationCrop-weed classification
MLP: Multi-Layer Perceptron, PNN: Probabilistic N Network, SOM: Self-Organizing Map, LVQ: Learning Vector Qantization, LLM: Local Linear Mapping
Comparison between methods
Katholieke Universiteit Leuven
• The strongest point is the local representation of the data accompanied by a local updating algorithm • Local updating algorithms assure much faster convergence than global updating algorithms (e.g. backpropagation for MLPs) • Because of the topologically preserving character of the SOM, the proposed classification method can deal with missing or noisy data, outperforming “optimal” classifiers (PNN) • The proposed method has been tested and gave superior results compared to a variety statistical and neural classifiers
Crop-weed classificationCrop-weed classification
Conclusions on LLM SOM technique
Precision spraying through controlled dose application
Unwanted variations in dose caused by horizontal and vertical boom movements
Precision treatment Precision treatment
Katholieke Universiteit Leuven
Active horizontal stabilisation of spray boom
Validation with ISO 5008 trackmovement of spray boom tip with and without controller
Indoor test of on-line weed detection and treatmentIndoor test of on-line weed detection and treatment
Precision treatment Precision treatment
Katholieke Universiteit Leuven
Sensor: Spectral line cameraClassification: Probabilistic neural networkProgram in Labview with c-code
Image acquisition frequence: 10 images/sec, travel speed: 30cm/sec, segmentation with NDVI ( > 0.3)Off-line training of NN, On-line classification Decision to spray:> 20 weed pixels and > 35% of vegetation is weed
Spray boom with PWM nozzles and controller, provided by Teejet Technologies
Indoor test of on-line weed detection and treatment
Precision treatment Precision treatment
Indoor test of on-line weed detection and treatment
Color image and spectral image
Precision treatment Precision treatment
Katholieke Universiteit Leuven
Indoor test - ResultsComparison of nozzle activation with weed positions
Precision treatment Precision treatment
Katholieke Universiteit Leuven
Indoor test - Results
camera
nozzle
weed
Experimental set up - separate weed classes (4) did not improve crop-weed classification