One Intelligent Video Surveillance System Based on DM642

2011 ◽  
Vol 474-476 ◽  
pp. 392-397
Author(s):  
Zhi Wei Tang ◽  
Xi Xuan Wu
Keyword(s):  
Image Processing ◽  
High Performance ◽  
Large Scale ◽  
Complex Image ◽  
Intelligent Video Surveillance System ◽  
Image Processing Algorithms ◽  
Processing Algorithms ◽  
Intelligent Information ◽  
Hard Real Time ◽  
Video Input

This article introduces an intelligent surveillance distributed system based on TMS320DM642. The system platform has many functions, such as OSD (on screen display), analog video output, digital video output, Hard Disk, Ethernet and so on. DM64. User can set the rules via the management software. The video input from analog cameras and IP cameras can be processed by DM642 according to the rules. If any event happens which acts against the rules, alarm will be given. The system provides immediate, accurate and intelligent services for users. In order to realize the complex image processing algorithms on DM642, we optimize the algorithms based on DSP and propose a series of rapid image processing algorithms. The design of the project puts the emphasis on the feasibility of distributed high-performance processing from both hardware and software aspects, which may be easily applied to other large scale or hard real-time intelligent information processing.

scholarly journals Road Object Detection: A Comparative Study of Deep Learning-Based Algorithms

Electronics ◽  
2021 ◽  
Vol 10 (16) ◽  
pp. 1932
Author(s):  
Malik Haris ◽  
Adam Glowacz
Keyword(s):  
Image Processing ◽  
Deep Learning ◽  
Object Detection ◽  
Real Time ◽  
Large Scale ◽  
Single Shot ◽  
Automated Driving ◽  
Convolutional Network ◽  
Image Processing Algorithms ◽  
Processing Algorithms

Automated driving and vehicle safety systems need object detection. It is important that object detection be accurate overall and robust to weather and environmental conditions and run in real-time. As a consequence of this approach, they require image processing algorithms to inspect the contents of images. This article compares the accuracy of five major image processing algorithms: Region-based Fully Convolutional Network (R-FCN), Mask Region-based Convolutional Neural Networks (Mask R-CNN), Single Shot Multi-Box Detector (SSD), RetinaNet, and You Only Look Once v4 (YOLOv4). In this comparative analysis, we used a large-scale Berkeley Deep Drive (BDD100K) dataset. Their strengths and limitations are analyzed based on parameters such as accuracy (with/without occlusion and truncation), computation time, precision-recall curve. The comparison is given in this article helpful in understanding the pros and cons of standard deep learning-based algorithms while operating under real-time deployment restrictions. We conclude that the YOLOv4 outperforms accurately in detecting difficult road target objects under complex road scenarios and weather conditions in an identical testing environment.

High Performance Parallelized Centroid Estimation of Image Components for Full Field Measurements

2014 ◽  
Author(s):  
Athanasios Iliopoulos ◽  
John G. Michopoulos
Keyword(s):  
Image Processing ◽  
High Performance ◽  
Field Measurements ◽  
Image Features ◽  
Full Field ◽  
Video Frames ◽  
Centroid Estimation ◽  
Efficient Data ◽  
Image Processing Algorithms ◽  
Processing Algorithms

Full field measurement methods require digital image processing algorithms to accomplish centroid identification of components of the image of a deforming structure and track them through subsequent video frames in order to establish displacement and strain measurements. Unfortunately, these image processing algorithms are the most computationally expensive tasks performed in such methods. In this work we present a set of new algorithms that can be used to identify centroids of image features that are shown to be orders of magnitude faster than conventional algorithms. These algorithms are based on employing efficient data structures and algorithmic flows tailored to optimally fit in shared memory parallel architectures.

Otoconia perfect/imperfect crystals

1994 ◽  
Vol 52 ◽  
pp. 42-43
Author(s):  
César D. Fermin ◽  
Dale Martin
Keyword(s):  
Image Processing ◽  
Room Temperature ◽  
Phosphotungstic Acid ◽  
Gallus Domesticus ◽  
Crystal Matrix ◽  
Organic Templates ◽  
Image Processing Algorithms ◽  
Processing Algorithms ◽  
Diffraction Patterns

Otoconia of higher vertebrates are interesting biological crystals that display the diffraction patterns of perfect crystals (e.g., calcite for birds and mammal) when intact, but fail to produce a regular crystallographic pattern when fixed. Image processing of the fixed crystal matrix, which resembles the organic templates of teeth and bone, failed to clarify a paradox of biomineralization described by Mann. Recently, we suggested that inner ear otoconia crystals contain growth plates that run in different directions, and that the arrangement of the plates may contribute to the turning angles seen at the hexagonal faces of the crystals.Using image processing algorithms described earlier, and Fourier Transform function (2FFT) of BioScan Optimas®, we evaluated the patterns in the packing of the otoconia fibrils of newly hatched chicks (Gallus domesticus) inner ears. Animals were fixed in situ by perfusion of 1% phosphotungstic acid (PTA) at room temperature through the left ventricle, after intraperitoneal Nembutal (35mg/Kg) deep anesthesia. Negatives were made with a Hitachi H-7100 TEM at 50K-400K magnifications. The negatives were then placed on a light box, where images were filtered and transferred to a 35 mm camera as described.

Sign in / Sign up

Export Citation Format

Share Document