Night Vision Imagery Research Articles

Automatic water level monitoring system based on computer vision technology for supporting the irrigation modernization

Measurement of water level in the irrigation channel is required to calculate the irrigation discharge so that crops can be irrigated precisely. Large irrigation areas require a real-time and accurate monitoring system that cuts operational costs. In a previous study, ultrasonic technology has already been developed and implemented. However, during the actual operation, visual verification is needed. To address the problem, an automatic water level monitoring system based on visual appearance using computer vision technology is developed. The image-based system works by automatically capturing canal gauge, identifying canal gauge in an image using colour detection method, measuring the length of the pixel of canal gauge, and convert it to an actual water level. As a result of the performance test, five photos were taken in 24 hours. One image was accurately measured by the microcomputer. This is because the intensity of sunlight affects the colour of the canal gauge so that the microcomputer incorrectly or inaccurately identifies the canal gauge length. A pre-calibration is needed to find a specific colour range for specific light intensity. Also a robust, less colour sensitive such as night vision imagery needs to be developed to overcome luminescence problems.

IMPLEMENTATION OF A REAL-TIME STACKING ALGORITHM IN A PHOTOGRAMMETRIC DIGITAL CAMERA FOR UAVS

Abstract. In the recent years, unmanned aerial vehicles (UAVs) have become an interesting tool in aerial photography and photogrammetry activities. In this context, some applications (like cloudy sky surveys, narrow-spectral imagery and night-vision imagery) need a longexposure time where one of the main problems is the motion blur caused by the erratic camera movements during image acquisition. This paper describes an automatic real-time stacking algorithm which produces a high photogrammetric quality final composite image with an equivalent long-exposure time using several images acquired with short-exposure times. Our method is inspired by feature-based image registration technique. The algorithm is implemented on the light-weight IGN camera, which has an IMU sensor and a SoC/FPGA. To obtain the correct parameters for the resampling of images, the presented method accurately estimates the geometrical relation between the first and the Nth image, taking into account the internal parameters and the distortion of the camera. Features are detected in the first image by the FAST detector, than homologous points on other images are obtained by template matching aided by the IMU sensors. The SoC/FPGA in the camera is used to speed up time-consuming parts of the algorithm such as features detection and images resampling in order to achieve a real-time performance as we want to write only the resulting final image to save bandwidth on the storage device. The paper includes a detailed description of the implemented algorithm, resource usage summary, resulting processing time, resulting images, as well as block diagrams of the described architecture. The resulting stacked image obtained on real surveys doesn’t seem visually impaired. Timing results demonstrate that our algorithm can be used in real-time since its processing time is less than the writing time of an image in the storage device. An interesting by-product of this algorithm is the 3D rotation estimated by a photogrammetric method between poses, which can be used to recalibrate in real-time the gyrometers of the IMU.

FACE: Fully Automated Context Enhancement for night-time video sequences

Unfavorable lighting conditions pose significant challenges to imaging devices. Use of multiple sensor fusion and false colorization has been demonstrated by many recent works. However, semi-automatic nature of these approaches make them less attractive. To address these shortcomings, we present a color night vision system, named FACE (Fully Automated Context Enhancement). It uses enhanced multi-spectral video fusion and fully automated color transform based color morphing to process multiple video streams captured by infrared (IR) and low light visible (VIS) sensors. At first, we introduce glare-suppressed inter-channel fusion in RGB color space to negate the ghost-like effect produced by IR channel. We then introduce color value imputation with deep KNN framework for joint classification of grayscale and color imagery. A suitable color transform is then sought to give a day-like color appearance to night vision imagery. Objective quality evaluation indicates the effectiveness of our framework for context enhancement at nighttime.

Fast colorization for single-band thermal video sequences

The color night-vision technology is advantageous to scene interpretation by displaying original monochrome night-vision imagery with colors, leading the future trend of night-vision development. Until recently, the methods of color night-vision are mostly based on multi-band sensor fusion that do not apply to thermal systems which are only sensitive to one spectral band. To address this, we present a novel technique in this paper aiming to directly give thermal videos day-time color appearance. We observe that the interframe pixel׳s luminance changing over thermal videos approximately follows a Double Lomax distribution. Based on this observation and computational reasons, we propose a star-shape mask for pixel matching. We further show how the mask can be used for efficient thermal video colorization. Additionally, considering the accumulating errors in sequence, we propose a simple region-similarity based abnormal color correction mechanism. Experiments on real applied thermal night-vision videos show that our method can efficiently render single-band thermal sequence with correct colors at 12–33 frame-per-second (fps), providing observers with enhanced facility of surrounding recognition and target recognition.

Augmenting full colour-fused multi-band night vision imagery with synthetic imagery in real-time

We present the design and first field trial results of an all-day all-weather enhanced and synthetic-fused multi-band colour night vision surveillance and observation system. The system augments a fused and dynamic three-band natural-colour night vision image with synthetic 3D imagery in real-time. The night vision sensor suite consists of three cameras, sensitive in, respectively, the visual (400–700 nm), the near-infrared (NIR, 700–1000 nm) and the long-wave infrared (LWIR, 8–14 µm) bands of the electromagnetic spectrum. The optical axes of the three cameras are aligned. Image quality of the fused sensor signals is enhanced in real-time through dynamic noise reduction, super resolution and local adaptive contrast enhancement. The quality of the LWIR image is enhanced through scene-based non-uniformity correction. The visual and NIR signals are used to represent the fused multi-band night vision image in natural daytime colours, using the Colour-the-Night colour remapping technique. Colour remapping can also be deployed to enhance the visibility of thermal targets that are camouflaged in the visual and NIR range of the spectrum. The dynamic false-colour night-time images are augmented with corresponding synthetic 3D scene views, generated in real-time using a geometric 3D scene model in combination with position and orientation information supplied by the Global Positioning System and inertial sensors of the system. Initial field trials show that this system provides enhanced situational information in various low-visibility conditions.

Towards an Understanding of the Effect of Night Vision Display Imagery on Scene Recognition

An experimental study was carried out to evaluate the combined effect of the absence of colour and the alteration of surface intensities on the recognition of complex scenes. This approach was motivated by the properties of the monochrome imagery of night vision goggles (NVGs), which are a commonly used form of night vision display. Observers were presented with pairs of aerial views of simulated urban scenes (from 400 or 700 ft), taken from viewing angles that differed by 30 deg to ensure that participants could not rely on a viewpoint-dependent cognitive representation of the scene. The observer's task was to decide whether the two scenes were the same, apart from the rotated viewpoint. On catch trials, which were fewer in number, one of the scenes was also mirror reversed. These trials were included to prevent guessing or premature responses. On half the trials, one of the scenes was rendered to simulate the effects of night-vision imagery. The time taken for the observers to confirm the identity of the rotated scenes was measured. There was no effect of differing altitude. When both of the scenes were rendered as daylight imagery, the average time to achieve a match was 34.7 s. When one of the scenes was rendered as NVD-style imagery, the matching time rose to 50.8 s. This effect varied according to the complexity of the scene and the extent to which the scene contained salient coloured features. There were also pronounced differences between observers. These findings suggest that night vision displays may have adverse effects on scene recognition compared to viewing natural-coloured images of the same scenes.