Artificial neural networks that would utilize the cascade error projection (CEP) algorithm have been proposed as means of autonomous, real-time, adaptive color segmentation of images that change with time.In the original intended application, such a neural network would be used to analyze digitized color video images of terrain on a remote planet as viewed from an uninhabited spacecraft approaching the planet.During descent toward the surface of the planet, information on the segmentation of the images into differently colored areas would be updated adaptively in real time to capture changes in contrast, brightness, and resolution, all in an effort to identify a safe and scientifically productive landing site and provide control feedback to steer the spacecraft toward that site. Potential terrestrial applications include monitoring images of crops to detect insect invasions and monitoring of buildings and other facilities to detect intruders.
The CEP algorithm is reliable and is well suited to implementation in very-large-scale integrated (VLSI) circuitry. It was chosen over other neural-network learning algorithms because it is better suited to real-time learning: It provides a self-evolving neural-network structure, requires fewer iterations to converge and is more tolerant to low resolution (that is, fewer bits) in the quantization of neural-network synaptic weights. Consequently, a CEP neural network learns relatively quickly, and the circuitry needed to implement it is relatively simple.
Like other neural networks, a CEP neural network includes an input layer, hidden units, and output units (see figure). As in other neural networks, a CEP network is presented with a succession of input training patterns, giving rise to a set of outputs that are compared with the desired outputs. Also as in other neural networks, the synaptic weights are updated iteratively in an effort to bring the outputs closer to target values. A distinctive feature of the CEP neural network and algorithm is that each update of synaptic weights takes place in conjunction with the addition of another hidden unit, which then remains in place as still other hidden units are added on subsequent iterations. For a given training pattern, the synaptic weight between (1) the inputs and the previously added hidden units and (2) the newly added hidden unit is updated by an amount proportional to the partial derivative of a quadratic error function with respect to the synaptic weight. The synaptic weight between the newly added hidden unit and each output unit is given by a more complex function that involves the errors between the outputs and their target values, the transfer functions (hyperbolic tangents) of the neural units, and the derivatives of the transfer functions.
The adaptive color-segmentation process of a proposed CEP can be summarized as follows: The knowledge acquired by the network up to a given time, t0, would be used in segmenting the image at the next increment of time, t0+Δt . The results of the segmentation at t0+Δt would then be used to update the knowledge pertaining to time t0. This segmentation and updating would be performed repeatedly as new imagery was acquired.
On the basis of (1) computational simulations using representative terrain images and (2) the performances of prior CEP integrated circuits, it has been estimated that adaptive learning can be achieved in times of the order of milliseconds. An important issue that must be addressed in practical development is how often updates must be performed: The frequency of updates would directly affect the power demand of the proposed CEP circuitry.
This work was done by Tuan A. Duong of Caltech for NASA's Jet Propulsion Laboratory. For further information, access the Technical Support Package (TSP) free on-line at www.techbriefs.com/tsp under the Information Sciences category. NPO-30692
This Brief includes a Technical Support Package (TSP).
Real-Time Adaptive Color Segmentation by Neural Networks
(reference NPO-30692) is currently available for download from the TSP library.
Don't have an account?
Overview
The document titled "Real-Time Adaptive Color Segmentation by Neural Networks" from NASA's Jet Propulsion Laboratory discusses advancements in color segmentation techniques essential for safe spacecraft landings. Traditional landing methods rely on pre-selected sites, which may not account for dynamic changes in terrain or lighting conditions. This lack of adaptability can lead to risks during descent, as spacecraft may encounter unrecognized hazards.
To address these challenges, the paper presents a new learning architecture that enables real-time adaptation of color segmentation as a lander approaches its landing site. The approach utilizes iterated learning from a sequence of images captured under varying light conditions, simulating the dynamic environment encountered during an actual descent. The study employs images obtained from JPL’s MARS YARD, taken at different times of the day, to evaluate the adaptive capabilities of the neural network.
The adaptive architecture allows the network to update its knowledge continuously, improving its ability to distinguish between different surface types, such as rocks and flat surfaces. This real-time adaptation is crucial for identifying safe landing zones, as it enables the system to respond to changes in light intensity, contrast, and resolution that occur during descent.
The document also emphasizes the importance of determining the optimal frequency for updates to the neural network. Frequent updates can lead to increased power consumption, while infrequent updates may result in insufficient training data for accurate segmentation. The balance between these factors is critical for the design of an effective spacecraft landing system.
Simulation results indicate that the adaptive color segmentation approach significantly enhances the accuracy of identifying safe landing zones compared to non-adaptive methods. The study concludes that real-time adaptive techniques are necessary for improving the safety and precision of spacecraft landings, paving the way for more ambitious exploration missions.
Overall, this research represents a significant step toward developing intelligent autonomous systems capable of making real-time decisions during critical phases of space missions, ultimately expanding the range of scientifically interesting landing sites for future exploration.
