InvestorsHub Logo
Followers 2
Posts 1540
Boards Moderated 0
Alias Born 05/06/2017

Re: None

Monday, 01/09/2023 12:43:11 PM

Monday, January 09, 2023 12:43:11 PM

Post# of 766
WiMi Designs Holographic AI Image Recognition-Based Obstacle Edge Detection Software.

Source
https://www.newstrail.com/wimi-designs-holographic-ai-image-recognition-based-obstacle-edge-detection-software/

December 4, 2022

As a large-capacity, high-density public transportation, the safety of urban rail transit is directly related to the safety of passengers’ lives. To ensure this system’s safe and efficient operation, a safe and reliable operation control system is necessary. Driverless rail transit is a transportation mode that can achieve full automation without driver intervention. With the progress of science and technology, unmanned and automated has become the development trend in the field of transportation, and the application of unmanned technology in China’s railroad construction is becoming more and more extensive. In the automatic traveling urban rail transit, the automated obstacle detection system is an essential function of the intelligent train, which can replace the driver’s observation of the obstacles on the track. Once the invasion of foreign objects is found, the control center and the driver can be notified in real-time, and the abnormal type and distance can be detected simultaneously to improve the safety and reliability of operation and reduce the operation and maintenance costs.

Active obstacle detection system uses a different technology than traditional train control systems. It applies the environmental perception technology of the automotive field. The sensors used are chosen to be consistent with the LIDAR, millimeter-wave radar, and camera used in automotive autonomous driving. And the machine learning algorithm used in automotive autonomous driving is used in the perception algorithm for target detection and classification. Multi-sensor fusion technology of the automotive field is also applied in the system.

WiMi Hologram Cloud, Inc. (NASDAQ:WIMI) R&D team developed unmanned rail vehicle obstacle edge detection software based on holographic image recognition technology, which is used to improve the recognition efficiency of track obstacle detection.WiMi’s R&D team added artificial intelligence deep learning technology to the obstacle detection software. Using a combination of long-distance track and near-distance obstacles helps increase the dataset of artificial intelligence image recognition technology for improving the detection function of long-distance objects of unmanned rail vehicle obstacle detection system and realizes the part of a more intelligent automatic judgment of obstacles and distance in actual operation.

Detecting tracks at long distances is more complicated than short-range detection:
the complex shape of the tracks hinders the pre-calculated track model, and the resolution is so low that it is difficult to distinguish the track edges from the structures of neighboring objects. WiMi’s R&D team uses an iterative analysis algorithm to expand the image detection sensor’s distance gradually. The position is calculated by iteration, which is used as the new position of the next target, and the further distance target is constantly obtained. The results obtained from a small range are used as the starting point for the next partition. In each iteration, the detected trajectory is projected onto a new sub-region, and then the corresponding matching ratio score is calculated based on the change of that sub-region.

In the future, WiMi’s R&D team will also optimize the unmanned track obstacle recognition system in several aspects. This would include optimizing the long-tail distribution of the dataset, enhancing the generalization capability of the dataset, adding more detection scenarios and obstacle types, optimizing the small target detection problem, and using cross-domain datasets to verify the generalization performance of the model, adopting multimodal detection methods, designing rail transit obstacle detection models based on computer vision, LIDAR, millimeter-wave radar and other multi-sensor fusion technologies, etc. With the development of driverless technology, new rail transportation methods will directly affect our future travel and bring more convenient new means of transportation for people’s life.
Volume:
Day Range:
Bid:
Ask:
Last Trade Time:
Total Trades:
  • 1D
  • 1M
  • 3M
  • 6M
  • 1Y
  • 5Y
Recent WIMI News