The software consists of only two panels realized in Visual Basic environmet with many visualization boxes, text boxes and buttons.

The 'Main panel' performs variaty of calculations for object recognition. The second screen is only for standard filtering operations like Roberts filter  and Friman language based recognition and analyzes. With our second panel we can compare between standard and proposed calculation. The software can function in automatic regime and has a possibilities to show the software reaction of object's calculation. This demo version can be transformed vary fast to real manufacturing system and tuning for special conditions like new details and levels of light intensity, some new camera view points and detail's geometry. It has a possibility to store data in database file and later some additional parameters to be analyzed. At the end you can manually perform and use some connections between the math functions into the system to be able to combine the best computational algorithms.

The increasing of flexibility of the supply in the automated assembling is subject to many research areas, one of which is the application of industrial robots equipped with computer vision systems. Such systems afford high flexibility, but there are still economical and technical obstacles for their extended application. This investigation aimed at constructing of a technical vision system in order to increase the flexibility of an assembling unit served by a robot. The system should also be economically effective and easy for application.

The possibilities for increasing the flexibility of the robotized unit were systematized through analyzing the subsystems of the unit and their share to the flexibility. An assembling unit was chosen from an available system consisting of a special conveyor, carriers and industrial robot SCARA. For scene of the computer vision system was chosen the carrier, which had proper surface, where the assembling details were recognized.

The basic variants for the position of the technical vision apparatus were formed. The optimal optical conditions and the performance of the system were proved through experiments.

The suggested model for increasing the flexibility of the robotized assembling unit has the following advantages: uniformity of the equipment not requiring special attachments, small changes in the kinematics of the position (in finding the proper place of the computer vision system), possibility for control of the robot hand, possibility for analysis and qualitative control of the surfaces of the details, possibility for processing of two- and three-dimensional surfaces, possibility for dynamic control of the position and relatively cheap components of the whole system.

Some experimental Results

Description	Real object	space rotation	object segmentation
	Applying nose

Here I can show you some real object calculated with my software

- On the each first picture you can see the object taken with fhree view point [angle between 30-80 degree from frontal direction].
- Each second picture represents scene (somme part with important for calculation information with objects). The computer represent this scene with frontal view point calculating basic picture point by point  transforming angle of basic plane.
- The third picture represents only the objject taken after some filters and algorithm for separating objects. This algorithm takes objects one by one from scene giving the best conditions to neuronet to calculate only one object in the moment. In this way you have a possibilities to calculate many objects in one scene.

I show you seven objects because the neuronet was trained for this objects.

Object num.	Real picture, segmentation and object recognition process
1		Object 1 is а driving wheel with diameter 9 sm. and scene is 10sm sq. This wheel has altitude 20mm.
2		Object 2 The middle picture represents a special filter (software for plane light intensity regulation)
3		Object 3
4		Object 4 This object has light reflection similar to the background (this is a synthetic material) and it is very difficult for recognition
5		Object 5
6		Object 6 Separate from 2 details
7		Object 7