Abstract: Detecting an object in an image of pixels is disclosed. The method includes: obtaining a position of a mask covering at least one pixel of an image; obtaining a predefined neighbourhood of pixels around the mask and dividing this neighbourhood into zones; for each zone, calculating a mean and variance of values of the pixels of the zone; calculating, for each possible pair of zones, a difference between the zones; determining whether a background of the mask is multimode or monomode according to the differences; calculating a variance of the values of the neighbourhood pixels using a formula for calculating the variance dependent on the monomode or multimode type of the background; calculating a signal to noise ratio for the mask position using the variance of the values of the pixels of the neighbourhood obtained; and detecting an object when the signal to noise ratio is above a threshold.

Abstract: Detecting an object in an image of pixels is disclosed. The method includes: obtaining a position of a mask covering at least one pixel of an image; obtaining a predefined neighbourhood of pixels around the mask and dividing this neighbourhood into zones; for each zone, calculating a mean and variance of values of the pixels of the zone; calculating, for each possible pair of zones, a difference between the zones; determining whether a background of the mask is multimode or monomode according to the differences; calculating a variance of the values of the neighbourhood pixels using a formula for calculating the variance dependent on the monomode or multimode type of the background; calculating a signal to noise ratio for the mask position using the variance of the values of the pixels of the neighbourhood obtained; and detecting an object when the signal to noise ratio is above a threshold.

Abstract: A method is provided for detecting and tracking targets in a series of images, at least one track as the existing track, being formed from spots detected in images in the series preceding a current image. The method includes, for the current image: obtaining at least one spot included in the current image and, for each spot, a value representing a characteristic of the spot; classifying each spot according to the representative value in a first category when the representative value is higher than a first predefined threshold or in a second category when the representative value lies between the first predefined threshold and a lower second predefined threshold; for each existing track, allocating a spot in the first category to the existing track and allocating a spot in the second category to the existing track when no spot in the first category is compatible with said existing track.

Abstract: A method is provided for detecting and tracking targets in a series of images, at least one track as the existing track, being formed from spots detected in images in the series preceding a current image. The method includes, for the current image: obtaining at least one spot included in the current image and, for each spot, a value representing a characteristic of the spot; classifying each spot according to the representative value in a first category when the representative value is higher than a first predefined threshold or in a second category when the representative value lies between the first predefined threshold and a lower second predefined threshold; for each existing track, allocating a spot in the first category to the existing track and allocating a spot in the second category to the existing track when no spot in the first category is compatible with said existing track.

Abstract: A defective-pixel detection method included in an image-processing procedure, including a pixel-processing procedure applied to pixels of an image supplied by an image sensor. Each pixel is associated with a classification value representing a state of said pixel. The method includes, for each pixel: applying the pixel-processing procedure; analyzing a result of the pixel-processing procedure; in the event of obtaining an unusual result representing a defect on a photosite of the image sensor that supplied said pixel, incrementing a number of detections of an unusual result for said pixel; and associating said pixel with a classification value representing a defective pixel when said number reaches a first threshold.

Abstract: The invention relates to a method for detecting and tracking targets in a series of successive images, said method comprising limiting the number of spots which are the subject of simultaneous tracking or false leads. The operation of the tracking module is thereby improved without having to increase a detection threshold of said spots. The detection threshold can even be reduced, such that the detection haul is increased and the tracking of each target is more continuous, without the probability of false alarms itself being increased.

Abstract: A defective-pixel detection method included in an image-processing procedure, including a pixel-processing procedure applied to pixels of an image supplied by an image sensor. Each pixel is associated with a classification value representing a state of said pixel. The method includes, for each pixel: applying the pixel-processing procedure; analysing a result of the pixel-processing procedure; in the event of obtaining an unusual result representing a defect on a photosite of the image sensor that supplied said pixel, incrementing a number of detections of an unusual result for said pixel; and associating said pixel with a classification value representing a defective pixel when said number reaches a first threshold.

Abstract: The invention relates to a method for detecting and tracking targets in a series of successive images, said method comprising limiting the number of spots which are the subject of simultaneous tracking or false leads. The operation of the tracking module is thereby improved without having to increase a detection threshold of said spots. The detection threshold can even be reduced, such that the detection haul is increased and the tracking of each target is more continuous, without the probability of false alarms itself being increased.