SEGMENTAL SURFACES IDENTIFICATION OF SEPARATE OBJECTS ON 2D-IMAGE IN TASKS OF THEIR MASS QUANTITY CONDITION ASSESSMENT
Read the full article
For citation: Ostapov D.S., Usatikov S.V. Segmental surfaces identification of separate objects on 2D-image in tasks of their mass quantity condition assessment. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2017, vol. 17, no. 4, pp. 685–693 (in Russian). doi: 10.17586/2226-1494-2017-17-4-685-693
We consider software of identification and assessment subsystem for condition of segmental surfaces of mass quantity of single material objects (as general population). Proposed solution supplements functionality of computer vision system (SCV) for recognition of types and sorts of objects on 2D-image. We suggest the algorithms of quantitative and qualitative assessment of each single material object condition (the element of general population) sensitive to details of their visible surface based on Pearson's criterion. Efficiency of a subsystem on different images is shown. We consider the possibilities of time reduction of video picture processing in an operating mode of computer vision system due to the decrease in computing complexity of algorithms and the increase in productivity of consecutive calculations and the organization of parallel calculations. The accuracy was calculated as the ratio of the number of correctly identified computer vision system states of the segments to their total number. It was revealed that on the images from the inspection conveyor with fruit raw materials, the SCV has an accuracy of 98.5% on the given type of objects. On the objects in the form of aspirin tablets, system of computer vision has accuracy of 93.2%, and at the objects of rice grain mass it is 95.4%.