DESC:[0016] The preferred embodiment of the present invention will now be explained with reference to the accompanying drawings. It should be understood however that the disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms. The following description and drawings are not to be construed as limiting the invention and numerous specific details are described to provide a thorough understanding of the present invention, as the basis for the claims and as a basis for teaching one skilled in the art how to make and/or use the invention. However, in certain instances, well-known or conventional details are not described in order not to unnecessarily obscure the present invention in detail.

[0017] The preferred embodiment of the present invention discloses an image processing system and method for computing the geometric distance estimation of obstacles at sea using an image capturing module fitted on a ship. Said process is done in real-time in order to produce immediate results. The system comprises of an image capturing module configured to capture at least one frame of video or RGB image of the sea surface in front of ship at sea, a control unit comprising of an object detection module, configured to identify the presence of an obstacle from captured RGB images as bounding boxes with diagonal image coordinates ((i, j), (i’, j ‘)) using deep learning, and distance estimation module configured to compute the distance to obstacles found in the sea. The captured video/images are processed by a pre-processing module in real time to remove noise and anomalies, wherein the processing steps are stored in the memory unit.

[0018] In the preferred embodiment of the present invention, the image capturing module is a mono camera attached to a pole at the front of a ship, wherein said camera is pointed downwards at an angle ‘a’ with respect to the pole, towards the sea surface. As per the preferred embodiment of the present invention, a single mono camera may be employed to find the geometric distance to obstacles at sea. The image capturing unit records at least one video frame to detect the object, wherein it detects said object’s motion on sea surface from multiple video frames.

[0019] Referring to FIG.1, assuming sea or water body surface as a geometric plane surface, the geometric distance estimation module estimates the distance to every pixel on sea surface from a pole-mounted camera at the front of the ship, using parameters such as camera pointing angle (a) with respect to the pole, height of camera position from sea surface, horizontal and vertical resolution of camera, horizontal field of view (Hfov) and vertical field of view (Vfov) of camera. The combination of sensor size and focal length will change the angle of view and the field of view. The angle of view remains constant always for a particular sensor and lens.
[0020] As shown in FIG.2, the image capturing module, comprising of a mono camera, is attached to a pole wherein said camera is pointed downwards at an angle ‘a’ with respect to the pole to capture RGB images of the sea surface in front of ship at sea. The image capturing process is continuous throughout the voyage of the ship. The object detection module, operatively coupled to the image capturing module, recognizes the objects from RGB images so captured as bounding boxes by extracting the diagonal image co-ordinates((i,j), (i’,j‘)), as illustrated in FIG.3. The distance estimation module, operatively coupled to the object detection module as well as the image capturing unit, uses the lower co-ordinates (i,j) of bounding boxes to compute the distances to obstacles in sea.

[0021] As depicted in FIG.4, the mono camera is mounted at a height ‘H’ metres from the sea level, wherein said camera’s field of view covers a rectangular area of sea on a RGB level. Considering the sea surface as a geometric plane surface, the system computes the distance to every pixel on sea surface using the height of camera position from sea surface, resolution of camera, and camera parameters Hfov and Vfov.

[0022] Referring to FIG.1 or FIG.4, AC=Distance to pixel (i, j), wherein Distance AB = H / Cos (a + (Vfov/ Nv)* j)……..equation 1.
(Nh, Nv) = Total horizontal and total vertical pixel depth resolution respectively of the RGB image.
(i, j) =pixel co-ordinate with respect to origin of RGB image;
H = camera height from water surface;
Alpha = camera angle;
Hfov = Horizontal field of view; and Vfov = Vertical field of view;
(Hfov/Nh) * i = Angle to i; (Vfov/ Nv)* j = Angle to j.

AC = Distance to the pixel (i,j) = Distance AB / Cos ( (Hfov/Nh) * i) …. equation 2.

[0023] The Lower Co-ordinates (i,j) of bounding boxes and equations 1 and 2 employed to compute the distances to obstacles in sea. Said obstacles could be boats, marine living beings, or humans

[0024] Although the present invention has been described in connection with the preferred embodiments thereof with reference to the accompanying drawings, it is to be noted that various changes and modifications are possible and are apparent to those skilled in the art. Such changes and modifications are to be understood as included within the scope of the present invention unless they depart there from.
,CLAIMS:1. A system for determining the geometric distance of obstacles at sea from a ship, said system characterized by:
• image capturing unit mounted on a pole at a pre-defined height and angle a with respect to the pole from the sea level, at the front of the ship, wherein said unit is configured to capture video frames as well as RGB images of the sea surface;
• pre-processing module configured to process the captured video/images in real time to remove noise and anomalies;
• memory unit to store the captured video frames and RGB images;
• object detection module configured to recognize objects on sea surface from the image captured by the image capturing unit, wherein it recognizes said objects from RGB images so captured as bounding boxes by extracting the diagonal image co-ordinates((i,j), (i’,j‘)); and
• distance estimation module, operatively coupled to the object detection module as well as the image capturing unit, configured to compute the distances to obstacles in sea using the lower co-ordinates (i,j) of bounding boxes.

2. The system as claimed in claim 1, wherein the image capturing unit is a mono camera, wherein the field of view of the image capturing unit covers a rectangular area of sea on a RGB level.

3. The system as claimed in claim 1 computes the distance to every pixel on RGB image captured by the mono camera using the height of camera position from sea surface, camera pointing angle a with respect to the pole, resolution of camera, and the external camera parameters Hfov and Vfov.

4. A method to find the geometric distance of obstacles at sea from a ship, said method characterized by:

• capturing a rectangular area of sea on a RGB level by the image capturing unit;
• storing said image in the memory unit;
• processing said images by the pre-processing module in real time to remove noise and anomalies, wherein said processed images are stored in the memory;
• identifying the presence of obstacles in the captured RGB images as bounding boxes with diagonal image coordinates ((i, j), (i’, j ‘)) by the object detection module; and
• determining the distance to the obstacles from height of camera position from sea surface, camera pointing angle a, horizontal and vertical resolution of camera, and horizontal field of view Hfov and vertical field of view Vfov, and lower co-ordinates (i,j) of bounding boxes.

5. A method to find the geometric distance to every pixel of an image of the sea surface captured by the image capturing unit, said method characterized by:

• capturing a rectangular area of sea on a RGB level by the image capturing unit;
• storing said image in the memory unit;
• processing said images by the pre-processing module in real time to remove noise and anomalies, wherein said processed images are stored in the memory; and
• determining the distance to every pixel on the captured image from the camera using height of camera position from sea surface, camera pointing angle a, horizontal and vertical resolution of camera, and horizontal field of view Hfov and vertical field of view Vfov.

01/09/2024
Prasanth S
IN/PA 2954
Agent for the Applicant