Search Projects Here

8 Results Found

AEROPLANE AND CAR PARKING IDENTIFICATION Classification Of Remote Sensing Images Using EfficientNet-B3 CNN Model With Attention

Scene Classification Is A Highly Useful Task In Remote Sensing (RS) Applications. Many Efforts Have Been Made To Improve The Accuracy Of RS Scene Classification. Scene Classification Is A Challenging Problem, Especially For Large Datasets With Tens Of Thousands Of Images With A Large Number Of Classes And Taken Under Different Circumstances. One Problem That Is Observed In Scene Classification Is The Fact That For A Given Scene, Only One Part Of It Indicates Which Class It Belongs To, Whereas The Other Parts Are Either Irrelevant Or They Actually Tend To Belong To Another Class. To Address This Issue, This Paper Proposes A Deep Attention Convolutional Neural Network (CNN) For Scene Classification In Remote Sensing. CNN Models Use Successive Convolutional Layers To Learn Feature Maps From Larger And Larger Regions (or Receptive Fields) Of The Scene. The Attention Mechanism Computes A New Feature Map As A Weighted Average Of These Original Feature Maps. In Particular, We Propose A Solution, Named EfficientNet-B3-Attn-2, Based On The Pre-trained EfficientNet-B3 CNN Enhanced With An Attention Mechanism. A Dedicated Branch Is Added To Layer 262 Of The Network, To Compute The Required Weights. These Weights Are Learned Automatically By Training The Whole CNN Model End-to-end Using The Back Propagation Algorithm. In This Way, The Network Learns To Emphasize Important Regions Of The Scene And Suppress The Regions That Are Irrelevant To The Classification. We Tested The Proposed EfficientNet-B3-Attn-2 On Six Popular Remote Sensing Datasets, Namely UC Merced, KSA, OPTIMAL-31, RSSCN7, WHU-RS19, And AID Datasets, Showing Its Strong Capabilities In Classifying RS Scenes.

BALL SPEED MEASUREMENT R2 -CNN Fast Tiny Object Detection In Large-Scale Remote Sensing Images

Recently, The Convolutional Neural Network Has Brought Impressive Improvements For Object Detection. However, Detecting Tiny Objects In Large-scale Remote Sensing Images Still Remains Challenging. First, The Extreme Large Input Size Makes The Existing Object Detection Solutions Too Slow For Practical Use. Second, The Massive And Complex Backgrounds Cause Serious False Alarms. Moreover, The Ultra-tiny Objects Increase The Difficulty Of Accurate Detection. To Tackle These Problems, We Propose A Unified And Self-reinforced Network Called Remote Sensing Region-based Convolutional Neural Network (R2-CNN), Composing Of Backbone Tiny-Net, Intermediate Global Attention Block, And Final Classifier And Detector. Tiny-Net Is A Lightweight Residual Structure, Which Enables Fast And Powerful Features Extraction From Inputs. Global Attention Block Is Built Upon Tiny-Net To Inhibit False Positives. Classifier Is Then Used To Predict The Existence Of Target In Each Patch, And Detector Is Followed To Locate Them Accurately If Available. The Classifier And Detector Are Mutually Reinforced With End-to-end Training, Which Further Speed Up The Process And Avoid False Alarms. Effectiveness Of R2-CNN Is Validated On Hundreds Of GF-1 Images And GF-2 Images That Are 18 000×18 192 Pixels, 2.0-m Resolution, And 27 620 × 29 200 Pixels, 0.8-m Resolution, Respectively. Specifically, We Can Process A GF-1 Image In 29.4 S On Titian X Just With Single Thread. According To Our Knowledge, No Previous Solution Can Detect The Tiny Object On Such Huge Remote Sensing Images Gracefully. We Believe That It Is A Significant Step Toward Practical Real-time Remote Sensing Systems.

HAND GESTURE RECOGNITION Automated Hand Gesture Recognition Using A Deep Convolutional Neural Network Model

The Tremendous Growth In The Domain Of Deep Learning Has Helped In Achieving Breakthroughs In Computer Vision Applications Especially After Convolutional Neural Networks Coming Into The Picture. The Unique Architecture Of CNNs Allows It To Extract Relevant Information From The Input Images Without Any Hand-tuning. Today, With Such Powerful Models We Have Quite A Flexibility Build Technology That May Ameliorate Human Life. One Such Technique Can Be Used For Detecting And Understanding Various Human Gestures As It Would Make The Human-machine Communication Effective. This Could Make The Conventional Input Devices Like Touchscreens, Mouse Pad, And Keyboards Redundant. Also, It Is Considered As A Highly Secure Tech Compared To Other Devices. In This Paper, Hand Gesture Technology Along With Convolutional Neural Networks Has Been Discovered Followed By The Construction Of A Deep Convolutional Neural Network To Build A Hand Gesture Recognition Application.

INDOOR AND OUTDOOR PEOPLE TRACKING NEAR WALL Real Time Detection And Tracking Of Human Based On Image Processing With Laser Pointing

Real-time Detection And Tracking Of A Moving Human Task Are Sorts Of Challenging If The Camera Itself Is Moving Or Place, Also With Tracking And Detection; We Are Going To Point The Human Using A Laser That Is Present On The System Itself. This Paper Displays An Execution Of Real-time Detection And Tracking Of A Human With 360 Degree Rotating Camera And A Stationery Camera. Adaption Of Various Tracking Of Object Algorithms And Their Effect On Execution And Implementation Is Also Presented. The System Explained In This Paper Has A Camera, Which Is Connected To An Embedded System (raspberry Pi Board). This Embedded System Or Raspberry Pi Board Runs An Image Processing Algorithm, Which Identifies A Person And Then Keeps Tracking The Person As Long As It's Within The Sight Of The Camera. Otherwise In The Video, The Person Is Tracked When The Camera Is Fixed As Stationery. If The Object Is Being Moved, The Raspberry Pi Gives A Signal To The Stepper Motor To Rotate The Camera For 360 Degree Tracking And Also The Laser Points The Person Simultaneously.

OBJECT DETECTION Object Detection And Tracking For Community Surveillance Using Transfer Learning

Automation Of Video Surveillance Is Gaining Extensive Interest Recently, Considering The Public Security Issues. In Recent Times, A Systematic And Exact Detection Of An Object Is A Foremost Point In Computer Vision Technology. With The Unfolding Of Recent Deep Learning Techniques, The Precision Of Detecting An Object Has Increased Greatly Thereby Igniting The Interest In This Area To Large Extent. Also, With The Advent Of Automatic Driving Electric Cars, The Accurate Detection Of Objects Has Gained Phenomenal Importance. The Main Aim Is To Integrate The State Of-the-art Deep Learning Method On Pedestrian Object Detection In Real-time With Improved Accuracy. One Of The Crucial Problems In Deep Learning Is Using Computer Vision Techniques, Which Tend To Slow Down The Process With Trivial Performance. In This Work, An Improved Yolov3 Transfer Learning-based Deep Learning Technique Is Used For Object Detection. It Is Also Shown That This Approach Can Be Used For Solving The Problem Of Object Detection In A Sustained Manner Having The Ability To Further Separate Occluded Objects. Moreover, The Use Of This Approach Has Enhanced The Accuracy Of Object Detection. The Network Used Is Trained On A Challenging Data Set And The Output Obtained Is Fast And Precise Which Is Helpful For The Application That Requires Object Detection.

TRAFFIC CONTROL USING VIDEO PROCESSING Smart Traffic Light Scheduling In Smart City Using Image And Video Processing

The Growing Population And Increased Vehicles Lead To The Main Challenges In Urban Life. Therefore, The Role Of Traffic Management Will Save Time And Fuel Consumption And Reduce Environmental Pollution. In Recent Years, Internet Of Things (IoT) And Smart Cities Drive A New Field Of Intelligent Traffic Management. In This Paper, A New Method For Traffic Light Control Is Presented By Using The Combination Of IoT And Image And Video Processing Techniques. In The Proposed Models, Traffic Light Scheduling Is Determined Based On The Density And The Number Of Passing Vehicles. Moreover, It Is Implemented By Raspberry-Pi Board And OpenCV Tool. The Analytical And Experimental Results Indicate The Efficiency Provided By The Proposed Models In Intelligent Traffic Management.

VIOLATED ACTION USING HELMET AND NUMBER PLATE RECOGNISATION

Automatic Detection Of Helmet For Motorcyclists From Real-time Surveillance Videos Is A Rising Application In Computer Science. Object Detection And Classification Using Deep Learning Was Recently Well-known Over The Years. Researchers Used These Techniques For Solving Several Surveillance-related Problems. Several Deep Learning Models Adopt For Automatic Detection Of Helmet For Motorcyclists But They Cannot Achieve State-of-the-art Results Due To Different Difficulties Such As Low Resolution, Whether Conditions, Occlusion And Illumination Etc. In This Paper, We Proposed A Methodology For Surveillance Video’s Images That Automatically Detect The Helmet Wear By Motorcyclist Or Not And He Is Using Mobile Or Not. For This Purpose, We Used The Faster CNN Model. Mail Will Be Send To The Rules Violators

WEED DETECTION Weed Detection Methods Based On Computer Vision

This Research Has Been Based On The Use Of Precision Agriculture Tools For The Management Of Weeds In Crops. It Has Focused On The Creation Of An Image-processing Algorithm To Detect The Existence Of Weeds In A Specific Site Of Crops. The Main Objective Has Been To Obtain A Formula So That A Weed Detection System Can Be Developed Through Binary Classifications. The Initial Step Of Image Processing Is The Detection Of Green Plants In Order To Eliminate All The Soil In The Image, Reducing Information That Is Not Necessary. Then, It Has Focused On The Vegetation By Segmentation And Eliminating Unwanted Information Through Medium And Morphological Filters. Finally, A Labeling Of Objects Has Been Made In The Image So That Weed Detection Can Be Done Using A Threshold Based On The Area Of Detection. This Algorithm Establishes An Accurate Monitoring Of Weeds And Can Be Implemented In Automated Systems For The Eradication Of Weeds In Crops, Either Through The Use Of Automated Sprayers For Specific Site Or A Weed Cutting Mechanism. In Addition, It Increases The Performance Of Operational Processes In Crop Management, Reducing The Time Spent Searching For Weeds Throughout A Plot Of Land And Focusing Weed Removal Tasks On Specific Sites For Effective Control.