Browse Prior Art Database

Pipelined Convolver for Two Dimensional Images

IP.com Disclosure Number: IPCOM000075001D
Original Publication Date: 1971-Jul-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 38K

Publishing Venue

IBM

Related People

Bakis, R: AUTHOR [+3]

Abstract

This is a system for processing in parallel the calculation of two dimensional convolutions. The system applies local convolution operators, in order to analyze data representing an image developed by a satellite communication system. The convolution operators are utilized to compensate for optical and environmental errors present in a received data image.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 71% of the total text.

Page 1 of 2

Pipelined Convolver for Two Dimensional Images

This is a system for processing in parallel the calculation of two dimensional convolutions. The system applies local convolution operators, in order to analyze data representing an image developed by a satellite communication system. The convolution operators are utilized to compensate for optical and environmental errors present in a received data image.

In Fig. 1, there is shown a representative geometry of image points, having a radius of three and expressed within an n, m coordinate system. In order to compensate for the error present in the information as discussed above, the value of the video data at point A, is replaced by the expression

(Image Omitted)

where, a represents an operator coefficient over the entire array, m, n and x represents a group of binary bits, each group being the analog value of the video image at a particular m, n point.

In Fig. 2, there is shown a block diagram of a one dimensional filter operating on the n-rows of video data. Each delay 12 indicated compensates for the time required to scan each row in the m dimension. The adders 14 perform the required summation to achieve the value y(A). The operators 10 execute the required multiplication operation of the operator coefficient, a, with the video data input x for each m, n point in the image.

In Fig. 3, there is shown a more detailed structure of a single row operator. Multiplication within the operator is performed by a table looku...