Stemmer Imaging develops new approach to vehicle classification
German machine-vision specialist Stemmer Imaging GmbH is applying a well-established image-processing method for the first time to real-time vehicle recognition applications.
In the first application, the class of vehicle - for example, a motor-cycle, car, light van, bus or truck - is identified, while in the second, the vehicle manufacturer, such as Ford, Mercedes or BMW, is determined. Both methods can be linked to systems that read license plates. The analysis is based on an object recognition tool that uses statistical learning theories.
Extremely promising results have been obtained in some preliminary trials and the company now needs to work with a dedicated partner to ensure that the appropriate traffic monitoring criteria are met. Potential applications for the systems include congestion charge monitoring and other road toll applications, traffic surveys and monitoring secure car parks.
There are many methods of locating patterns in images, however no single approach is best for all applications, says Stemmer Imaging. According to the company, its pattern recognition tool CVB Manto (which is part of the Common Vision Blox hardware-independent machine-vision tool kit) combines the positive characteristics of existing tools, enabling completely new application areas to be addressed.
CVB Manto uses the following features of an image: correlation, geometrical connections, texture and colour. A non-linear multi-resolution filter (MRF) is used to independently transform the picture, determine the relevant features of an object and enter these into a feature vector. A neural network then separates the object classes on the basis of these features. Essentially, the system learns to identify the patterns of interest from a set of training images and then calculates a confidence factor for its classification choice for each test image.
For traditional neural networks, there is a finite number of training images that the system can deal with - a phenomenon known as 'over fitting'. When this point is reached, the effort to separate classes becomes too difficult and error rate saturation occurs. As the number of sample images increases, the error rate drops until it reaches a saturation point and does not improve any further. According to Stemmer Imaging, the support vector machine approach does not suffer from over fitting problems, allowing the use of any number of training images, providing sufficient memory can be allocated to the fitting process. As a result, CVB Manto is capable of classifying objects with an accuracy not previously possible, claims the company.
The automatic recognition and reading of road signs during a journey is an interesting research project on which almost every vehicle manufacturer is currently working. One of the main problems is that the appearance of signs is severely affected by weather conditions. CVB Manto provides a solution to this problem. Test drives have shown that image-processing systems based on this pattern recognition tool can reliably recognise road signs that have been learnt. On this basis, drivers in the future could, for instance, receive warning if the current speed limit has been exceeded, or if they are intending to drive the wrong way down a one-way street.
CVB Manto also opens up a whole range of other possibilities, such as identifying pedestrians or cyclists, maintaining a differential distance from the vehicle in front or tracking lane markings, a prerequisite in the development of autonomous vehicles.
Contact:
Stemmer Imaging GmbH
Gutenbergstrasse 9-11
82178 Puchheim
Germany
Phone: +49-89-80902-0
Fax: +49-89-80902-116
Email: info@imaging.de
Web: http://www.stemmer-imaging.de
Web: http://www.commonvisionblox.com
In the first application, the class of vehicle - for example, a motor-cycle, car, light van, bus or truck - is identified, while in the second, the vehicle manufacturer, such as Ford, Mercedes or BMW, is determined. Both methods can be linked to systems that read license plates. The analysis is based on an object recognition tool that uses statistical learning theories.
Extremely promising results have been obtained in some preliminary trials and the company now needs to work with a dedicated partner to ensure that the appropriate traffic monitoring criteria are met. Potential applications for the systems include congestion charge monitoring and other road toll applications, traffic surveys and monitoring secure car parks.
There are many methods of locating patterns in images, however no single approach is best for all applications, says Stemmer Imaging. According to the company, its pattern recognition tool CVB Manto (which is part of the Common Vision Blox hardware-independent machine-vision tool kit) combines the positive characteristics of existing tools, enabling completely new application areas to be addressed.
CVB Manto uses the following features of an image: correlation, geometrical connections, texture and colour. A non-linear multi-resolution filter (MRF) is used to independently transform the picture, determine the relevant features of an object and enter these into a feature vector. A neural network then separates the object classes on the basis of these features. Essentially, the system learns to identify the patterns of interest from a set of training images and then calculates a confidence factor for its classification choice for each test image.
For traditional neural networks, there is a finite number of training images that the system can deal with - a phenomenon known as 'over fitting'. When this point is reached, the effort to separate classes becomes too difficult and error rate saturation occurs. As the number of sample images increases, the error rate drops until it reaches a saturation point and does not improve any further. According to Stemmer Imaging, the support vector machine approach does not suffer from over fitting problems, allowing the use of any number of training images, providing sufficient memory can be allocated to the fitting process. As a result, CVB Manto is capable of classifying objects with an accuracy not previously possible, claims the company.
The automatic recognition and reading of road signs during a journey is an interesting research project on which almost every vehicle manufacturer is currently working. One of the main problems is that the appearance of signs is severely affected by weather conditions. CVB Manto provides a solution to this problem. Test drives have shown that image-processing systems based on this pattern recognition tool can reliably recognise road signs that have been learnt. On this basis, drivers in the future could, for instance, receive warning if the current speed limit has been exceeded, or if they are intending to drive the wrong way down a one-way street.
CVB Manto also opens up a whole range of other possibilities, such as identifying pedestrians or cyclists, maintaining a differential distance from the vehicle in front or tracking lane markings, a prerequisite in the development of autonomous vehicles.
Contact:
Stemmer Imaging GmbH
Gutenbergstrasse 9-11
82178 Puchheim
Germany
Phone: +49-89-80902-0
Fax: +49-89-80902-116
Email: info@imaging.de
Web: http://www.stemmer-imaging.de
Web: http://www.commonvisionblox.com
