Table of Contents
Real time object detection and tracking
Real time object detection and tracking. Today’s security is given much significance in security applications. To monitor the continuous movements of persons or vehicles and reporting when predefined events take place is very common security applications. In this article a method is described for tracking moving objects from a sequence of video frame. This report mainly presents real time object tracking system. The foremost objective is to detect an object from the real time video and track the object. Tracking is done by using many techniques, Image Subtraction, Kalman Filter and Optical Flow. But here we will use optical flow method. Real time simulations are carried out using MATLAB (Image Acquisition and Image Processing Toolbox). This report finds many applications in surveillance, tracking radar, spying, monitoring and visual positioning and tracking.
The objective of this project is to identify and track a moving object within a real time video sequence. The detection and tracking of the object is based on Optical flow. The proposed method is straightforward and easier to implement. Tracking of a moving object in real time depends on the techniques of motion detection. Automated detection and tracking of moving objects in a live video has become an important research topic as well as the technology related to image processing due to the increasing popularity of surveillance and security systems. It deals with detecting moving objects such as humans or cars in digital images and video sequences. MATLAB is one of the tools used for the implementation of this algorithm.
Object detection has applications in many areas of computer vision, as well as image renewal and video surveillance. The results of detection, recognition and tracking has applications in many areas such as, for surveillance, tracking radar, automatic fire control, object counting and monitoring, automated vehicle system and visual positioning. It can be applied further to the areas of semantic indexing of video. Visual tracking system can be divided into two major parts: 1) detecting the target, and 2) filtering and association of data. Real-time segmentation of moving regions in image sequences is a fundamental step in many vision systems.
We will design a unmanned Arial vehicle that will be used for autonomous real time tracking system with the help of wireless camera mounted on it. Unmanned Arial vehicle have the ability to move swiftly and agilely through the air. Wireless camera sends an analog video which will be converted to a digital using analog to digital video converter. Computer will receive a live video and processing is done by using MATLAB which begins to track moving objects.
The methods proposed for tracking are Kalman Filter, Optical Flow and Image subtraction. A representative method for object detection is background subtraction. Background and foreground objects are needed to be separated, processed and analyzed. The data acquired from this is then used to promote it to motion detection. Shadows are also taken under consideration in real time problems while detecting motion.
MILESTONES:
Following are goals in the real time object detection and tracking:
- Develop a platform for UAV
- Extract the maximum advantage from the digital recorded data to detect motion
- Detect the real time moving object without human eye
- Monitor continuously movements all the time when pre-defined events take place and capture these events
SYSTEM DESCRIPTION:
Hardware requirements
DJI Phantom 2 H3-3D
The Phantom 2 is unbelievably easy to fly. Combined with smooth, stabilized footage from the H3-2D or H3-3D gimbal, aerial cinematography and videography has never been easier. It’s light, rugged and perfect for travel. Requires Zenmuse H3-2D or H3-3D, video downlink and monitor.
Features:
- Stabilization Gimbal: 3-axis Camera Stabilization Gimbal
- Video Downlink1:Max 700m Wi-Fi Connection
- APP:DJI VISION App
- Camera:14 Megapixels/1080p
- Controllable Range:-90°-0° (Vertical)
- Remote Control :800m (FCC), 5.8Ghz
preinstalled smartphone holder - Max Flight Time: 25 min.
Some other features are:
- Easy to install, Plug and Play
No additional manual adjustment required, no need to adjust the C.G. or parameters in the software. In addition, the installation process is extremely simple and the plug-and-play characteristics make it easy for users to experience the performance of the H3-2D.
- Customized for GoPro Hero3
The H3-2D gimbal is designed for the GoPro Hero3, and remote start/stop shooting, video transmission (video downlink required), etc of the GoPro Hero3 will be supported by future firmware upgrades.
- Compatible with DJI flight control systems
The H3-2D gimbal is compatible with DJI flight control systems, it can adjust the roll and pitch angle according to the accurate control of the Gimbal feedback and communicate with the MC in real-time.
It supports WooKong-M, NAZA-M V1, NAZA-M V2, Phantom;
It will support Ace One, WooKong-H, NAZA-H;
It doesn’t support NAZA-M LITE.
Notice: NAZA-M V1 and Phantom users need the PMU of NAZA-M V2.
- The Core Technology Inherited from the Zenmuse
The H3-2D gimbal inherited the core algorithm from the Zenmuse Z15 gimbal and the unmanned aircraft flying attitude stability control algorithm. Combined with the high precision sensors and brushless motors, the DJI Zenmuse H3-2D gimbal provides pixel level stabilized footage.
- Independent IMU module
The independent high-precision IMU module provides precise attitude data and the closed-loop control gives improved control accuracy.Without the requirement of complex installation and connection of many wires, it is also easy to assemble and remains stable when in flight.
- Extremely light weight with solid mechanical design
The aluminum alloy material and innovative mechanical design makes the gimbal extremely light weight (total 230g, GoPro Hero3 included). The structural design of the aluminum frame guarantees the strength of the gimbal.
- Ideal Choice for Phantom and Flame Wheel
The mechanical design, weight and power supply will match the Phantom and Flame Wheel perfectly. The Phantom transmitter can be easily upgraded to give the pitch(tilt) control function of the gimbal, with a customized slider. No additional receiver channels are needed, it just uses the additional MC channels.
- Online Upgrade
The functionality can be easily updated by online firmware upgrades
The DJI Lightbridge 2.4G full HD digital video downlink features a native OSD system that needs no additional equipment, for an all-in-one FPV experience. Once connected, you can control your flight controller through its built-in 2.4G remote control link or DJI 5.8G remote control.
2.3 ANALOG TO DIGITAL CONVERTER:
This is basically a TV box USB device. We use this device to convert our real time video i.e.,. output of the wireless camera into digital video. We convert our video into digital to process it by a digital computer. Wireless camera captures the video and is then received by the receiver. Receiver then sends these signals to analog to digital converter video converter which is then switched to digital cable.
Also read here