Image monitoring method, image monitoring apparatus and storage media

Abstract

In order to precisely detect abnormal objects by use of a low-cost arrangement at high speeds while avoiding unwanted influence of variations of a light source and / or regularly vibrating objects in the environment concerned, there are provided a means 4200 for subdividing an image being input from a camera into blocks, an object candidate extraction unit 4700 which is operable to compare image data of a frame to be processed and image data of its immediately preceding frame in units of blocks to thereby extract an abnormal object candidate in accordance with the presence or absence of edges and a longitude-to-lateral edge radio change rate, and an object judging unit 4800 for determining or “judging” whether the abnormal object candidate is a true abnormal object or not, wherein the object judging unit 4800 regards it as an abnormal object in cases where movement of the abnormal object candidate was traceable for a prespecified length of time period.

Description

BACKGROUND OF THE INVENTION [0001] The present invention relates in general to technologies for monitoring scenes to be monitored as captured or photographed by cameras in the events of surveillance in road transportation monitoring fields and private surveillance within buildings plus outdoor surveillance or the like. More particularly, but not exclusively, this invention relates to a method and apparatus for monitoring of images to detect abnormalities through image processing of image data of scenes being monitored. The invention also relates to storage media for storing therein computer programs used to realize the image monitoring methodology. [0002] As prior known image monitoring or surveillance apparatus, there is a method for performing abnormality detection through the steps of coding an image as captured or “photographed” by a camera into compressed data involving with-time differential image components and then directly utilizing such with-time differential images in thi...

AI Technical Summary

Benefits of technology

This approach enables accurate detection of true abnormalities while reducing processing time and costs, improving detection precision, and allowing for automatic tracking and communication of abnormal events.

Problems solved by technology

Unfortunately this method is encountered with a problem that under the circumstances with luminance variations with time such as brightness changes due to flickering of fluorescent lamps and / or turn-on / off of illumination devices and daylight variation of sunlight and clouds or the like, these can be detected as abnormalities incorrectly.

In addition, due to the fact that tree swing with regular vibration in a direction of movement or alternatively curtain movements or else causes the luminance to change with time, the use of this method would result in incorrect detection of them as abnormalities in some cases.

This method, however, suffers from problems which follow: The need for extension or “decompression” of such once-compressed image results in an unwanted increase in image information when compared to the case of compression; the resultant costs can increase undesirably due to an increase in length of time required for execution of the intended processing and also the necessity to employ an extra expansion device.

A further problem faced with the prior art method is that detection accuracy stays lower.

This can be said because the method tends to incorrectly detect the inherently non-abnormal objects such as tree swing with regular vibration in movement direction, curtain movement, or else as abnormal objects due to the fact that these exhibit changes in comparison with the background scene image thereof.

Images