Asaad F. Said

Automated Detection and Classification of Non-Wet Solder Joints

Non-wet solder joints in processor sockets are causing mother board failures. These board failures can escape to customers resulting in returns and dissatisfaction. The current process to identify these non-wets is to use a 2D or advanced X-ray tool with multidimension capability to image solder joints in processor sockets. The images are then examined by an operator who determines if each individual joint is good or bad. There can be an average of 150 images for an operator to examine for each socket. Each image contains more than 30 joints. These factors make the inspection process time consuming and the output variable depending on the skill and alertness of the operator. This paper presents an automatic defect identification and classification system for the detection of non-wet solder joints. The main components of the proposed system consist of region of interest (ROI) segmentation, feature extraction, reference-free classification, and automatic mapping. The ROI segmentation process is a noise-resilient segmentation method for the joint area. The centroids of the segmented joints (ROIs) are used as feature parameters to detect the suspect joints. The proposed reference-free classification can detect defective joints in the considered images with high accuracy without the need for training data or reference images. An automatic mapping procedure which maps the positions of all joints to a known Master Ball Grid Array file is used to get the precise label and location of the suspect joint for display to the operator and collection of non-wet statistics. The accuracy of the proposed system was determined to be 95.8% based on the examination of 56 sockets (76 496 joints). The false alarm rate is 1.1%. In comparison, the detection rate of a currently available advanced X-ray tool with multidimension capability is in the range of 43% to 75%. The proposed method reduces the operator effort to examine individual images by 89.6% (from looking at 154 images to 16 images) by presenting only images with suspect joints for inspection. When non-wet joints are missed, the presented system has been shown to identify the neighboring joints. This fact provides the operator with the capability to make 100% detection of all non-wets when utilizing a user interface that highlights the suspect joint area. The system works with a 2D X-ray imaging device, which saves cost over more expensive advanced X-ray tools with multidimension capability. The proposed scheme is relatively inexpensive to implement, easy to set up and can work with a variety of 2D X-ray tools.

Robust automatic void detection in solder balls

Voids in solder balls can cause board failures. The detection and assessment of voids in solder balls can help in reducing board yield issues caused by incorrect scrapping and rework. X-ray imaging machines make voids visible to the operator for manual inspection. Some existing x-ray inspection systems have void detection algorithms that require the use of intensive manual tuning operations that are time consuming, and inaccurate due to the inability to examine balls overshadowed with other components. In this paper, a robust automatic void detection algorithm is proposed. The proposed method is able to detect voids with different sizes inside the solder balls, including the ones that are overshadowed by board components and under different brightness conditions. Results show that the proposed method achieves a correlation squared in the range of 91% to 97 % with ground truth data from a 3D x-ray scan. The proposed algorithm is fully automated and benefits the manufacturing process by reducing operator effort and by providing a cost effective solution to improve output quality.

Multi-Region Texture Image Segmentation Based on Constrained Level-Set Evolution Functions

A multi-region texture image segmentation method based on level-set is proposed in this paper. In the proposed method, each region is represented by one level-set function and these functions evolve simultaneously based on a constraint. The constraint is used to keep a balance between competing regions and to guarantee disjoint and non-overlapping regions. To speed up the curve evolution functions and to prevent them from getting stuck at undesired points, a region competition factor is applied. Edge- and edgeless-based active contours are applied in the proposed method to improve the robustness and the accuracy of the segmentation. The proposed multi-region texture segmentation method is fast and less sensitive to initializations as compared with existing techniques. Different segmentation examples are presented to illustrate the performance of the proposed method.

White and Color Noise Cancellation using Adaptive Feedback Cross-Coupled Line Enhancer Filter

In this paper, a new method for white and color noise cancellation is presented. The proposed method consists of two adaptive FIR filters, where the output of each filter is feedback to the input of the other filter to form an adaptive feedback cross-coupled filter. The step-size parameter of each FIR filters is variable and based on power normalization techniques. The power normalization step is greatly improve the performance of the proposed filter and gives extra output SNR improvements compared to the existing methods. Illustrative examples are given in this paper to show the performance of the proposed method in removing white, color, or both white and color noises at very different input signal to noise ratios (SNRs).

Automated Void Detection in Solder Balls in the Presence of Vias and Other Artifacts

Voids are one of the major defects in solder balls and their detection and assessment can help in reducing unit and board yield issues caused by excessive or very large voids. Voids are difficult to detect using manual inspection alone. 2-D X-ray machines are often used to make voids visible to an operator for manual inspection. Automated methods do not give good accuracy in void detection and measurement because of a number of challenges present in 2-D X-ray images. Some of these challenges include vias, plated-through holes, reflections from the the plating or vias, inconsistent lighting, background traces, noise, void-like artifacts, and parallax effects. None of the existing methods that has been researched or utilized in equipment could accurately and repeatably detect voids in the presence of these challenges. This paper proposes a robust automatic void detection algorithm that detects voids accurately and repeatably in the presence of the aforementioned challenges. The proposed method operates on the 2-D X-ray images by first segregating each individual solder ball, including balls that are overshadowed by components, in preparation for treating each ball independently for void detection. Feature parameters are extracted through different classification steps to classify each artifact detected inside the solder ball as a candidate or phantom void. Several classification steps are used to tackle the challenges exhibited in the 2-D X-ray images. The proposed method is able to detect different-sized voids inside the solder balls under different brightness conditions and voids that are partially obscured by vias. Results show that the proposed method achieves a correlation squared of 86% when compared with manually measured and averaged data from experienced operators from both 2-D and 3-D X-ray tools. The proposed algorithm is fully automated and benefits the manufacturing process by reducing operator inspection time and removing the manual measurement variability from the results, thus providing a cost-effective solution to improve output product quality.

Non-wet solder joint detection in processor sockets and BGA assemblies

The existence of non-wet solder joints in PCB sockets can cause boards failures and its necessary to inspect theses sockets to locate any possible defective joints. 2D or advanced x-ray machines are used to image solder joints in processor sockets and make solder joints visible to be examined by the operator who determines if each individual joint is defective or not. This is a very time consuming process since each processor has an average of 150 images with 30 joints per image. An accurate and efficient non-wet detection method is proposed in this paper. The main components of the proposed method consist of region of interest (ROI) segmentation, feature extraction, reference-free classification, and automatic mapping. The ROI segmentation process is a noise-resilient segmentation method for the joint area. The centroids of the segmented joints (ROIs) are used as feature parameters to detect the suspect joints. The proposed reference-free classification can detect defective joints with high accuracy without the need for training data. An automatic mapping method is used to get the precise label and location of the suspect joint. The accuracy of the proposed method was determined to be 95.8% detection rate with 1.1% false alarm rate based on the examination of 56 sockets (500K joints). In comparison, the detection rates of currently available advanced x-ray tools with multi-dimension capability are in the range of 43% to 75%. The proposed method reduces the operator effort by 90%. The presented system identifies neighboring joints to any missed non-wet joints, which provides an operator with the capability to make 100% detection of all non-wets. The proposed scheme works with a 2D x-ray imaging device, which makes the proposed scheme relatively inexpensive to implement, and is very portable and easy to set up as compared to the available advanced x-ray tools with multi-dimension capability.

Cell Migration Analysis using a Statistical Level-Set Segmentation on a Wavelet-Based Structure Tensor Feature Space

In this paper, a new noise-resilient cell migration analysis scheme for bladder cancer cells is presented. The proposed scheme is based on texture segmentation in the wavelet domain using structure tensor features and adaptive statistical level-set segmentation. The proposed method extracts the region of interest where the cells are clustering, at different time instances, and computes the overall migration cell rate. For this purpose, the structure tensor data is processed using a trous wavelet filtering, which speeds up the algorithm as compared to existing nonlinear diffusion filters with the same accuracy. The proposed scheme is robust to noise and natural artifacts in the bladder cancer cell images. Moreover, the scheme can be applied successfully to images with poor contrast and high cell concentrations, even when the cells are overlapping and tiny. Simulation results are presented to show the performance of the proposed scheme.

MIGRATION AND PROLIFERATION ANALYSIS FOR BLADDER CANCER CELLS

This paper presents a cell evolution analysis (CEA) scheme for bladder cancer cells. The proposed scheme consists of a cell migration analysis component for computing the overall migration rate of the cell cluster, and a cell proliferation analysis component based on counting the individually segmented cells within the cell cluster at different time points. The proposed CEA scheme performs well for images with poor contrast and high cell concentrations, even when the cells are overlapping and small. Results are presented to illustrate the performance of the proposed scheme

Die level defects detection in semiconductor units

The assembly test process has many steps where defects can be created at any time during different stages. Detecting defects at early stages is very crucial and saves a lot of cost and time by isolating the defective parts from further processing. Detecting defects on the die area of the semiconductor units is a challenging procedure due to the fact that dies defects exhibits large variations in intensity and shape. The existing manual and automated inspection approaches still produce high rate of under-rejection and over-rejection which impacts the yield and adds significant cost to the inspected unit. A robust die level defect detection procedure is presented in this paper in order to come up with a solution that is cheaper, easier to sustain, and that would automatically inspect each unit for defects providing for efficient baseline characterization and rapid excursion detection. The proposed method gives a higher detection rate of die level defects with under-rejecting and over-rejection rates within the acceptable criteria.

Real-time detection and classification of traffic light signals

Traffic light detection is an important part of Advanced Driver Assist as well as autonomous vehicle systems which ensures timely and appropriate reaction to traffic lights (TLs) in cross sections. In this paper we introduce a robust and realtime approach to detect TLs and recognize its status in complex traffic scenes solely based on image processing techniques. The proposed system uses color properties of the scene to detect TLs in real-time. An innovative technique has been developed to significantly decrease compute requirement for detection of TL color by using one Lookup Table independent of lighting conditions. Each candidate region is further analyzed, using features analysis, to segregate actual TL signals among all candidate regions. As in similar machine learning techniques, an unsupervised classifier using a set of significant features has been developed to accurately segregate circular, semi-circular, and arrow shaped TL signals without using a training dataset. The final C++ code has been implemented and optimized on intelplatform using 1920x1080 frame resolution to recognize the status of TLs during day-time and night-time scenes, achieving 95% precision and 94.7% recall at 30FPS.