← В ленту
Регистрация: 12.01.2024

Othmane Naggar

Специализация: Robotics Perception Engineer

Портфолио

MAScIR

Real-time traffic monitoring solution for vehicle counting speed estimation and traffic jam detection : • Collect of videos of different roads in morocco with vehicles. • Developpement of CNN algorithm to detect vehicles in images using (keras python). • Developpement of vehicles speed estimation using(opencv), background subtraction and stereoscopic effect to estimate distances then speed using (opencv)and geometry. • Development of a vehicle classifier in 3 categories using Yolov2 • Development of a the ntcip protocol in C++ to send traffic data to the client server.

MAScIR

Street light detection algorithms by satellite imagery: • Buying a night satellite image of the city of Rabat of JILIN-1 satellite. • Development of a Python script under (QGIS) software to match the positions of streetlights with the positions given by client. • Training and validation of a deep neural network (keraspython) to classify streetlight by ON/OFF depending on the shape of the hallow and pixels intensity.

MAScIR

Autonomous greenhouse (growth estimation and disease detection of salads) : • Choice of the spectral camera (4 bands RGB and NIR from Spectral Devices). • Development of the communication layer between the dashboard and the camera using the sdk in C++. • Implementation of the salad growth estimation module. • Implementation of the disease detection module using a public database and Yolov3.

Скиллы

C++
CNN
Computer Vision
Data Mining
Deep Learning
Keras
KNN
LSTM
Machine Learning
Matplotlib
Numpy
Opencv
Python
Pytorch
Scikit-learn
Tensorflow

Опыт работы

Machine learning engineer
с 10.2016 - По настоящий момент |MAScIR
CNN, Yolo, C++
Real-time traffic monitoring solution for vehicle counting speed estimation and traffic jam detection : • Collect of videos of different roads in morocco with vehicles. • Developpement of CNN algorithm to detect vehicles in images using (keras python). • Developpement of vehicles speed estimation using(opencv), background subtraction and stereoscopic effect to estimate distances then speed using (opencv)and geometry. • Development of a vehicle classifier in 3 categories using Yolov2 • Development of a the ntcip protocol in C++ to send traffic data to the client server.
Machine learning engineer
с 10.2016 - По настоящий момент |MAScIR
Python
Street light detection algorithms by satellite imagery: • Buying a night satellite image of the city of Rabat of JILIN-1 satellite. • Development of a Python script under (QGIS) software to match the positions of streetlights with the positions given by client. • Training and validation of a deep neural network (keraspython) to classify streetlight by ON/OFF depending on the shape of the hallow and pixels intensity.
Machine learning engineer
с 10.2016 - По настоящий момент |MAScIR
Python, Qt, Opencv, C++
Citrus counting on trees to estimate the yield of a plot : • Data collection of more than 5000 images of moroccan citrus trees in different regions of the country. • Development of an annotation tool for the tree images. • Training and validation of a citrus fruit detector using Yolo and imagenet transfer learning. • Training and validation of a crop yield estimator using as input the number of visible fruits on tree images and climate and irrigation informations using CNN and LSTM. • Development of graphical interface for the client using QT, C++.
Machine learning engineer
с 10.2016 - По настоящий момент |MAScIR
C++, Opencv, Yolov3
Autonomous greenhouse (growth estimation and disease detection of salads) : • Choice of the spectral camera (4 bands RGB and NIR from Spectral Devices). • Development of the communication layer between the dashboard and the camera using the sdk in C++. • Implementation of the salad growth estimation module. • Implementation of the disease detection module using a public database and Yolov3.
Machine learning engineer
с 10.2016 - По настоящий момент |MAScIR
C++, Qt, Opencv, Specim, Tensorflow, Keras, Python, Numpy, Pandas, CNN
Sterile phosphate classification by hyperspectral imaging: • Development of a desktop app for Hyperspectral Data Collection. • Training and validation of a deep neural network models to classify phosphate using CNNs and f1 score metric. • Drafting of a detailed study report containing the results of the study.

Образование

Specialization
Educational institution

Языки

АнглийскийРодной